❌

Reading view

There are new articles available, click to refresh the page.

The Paramount-Warner Brothers Deal Could Usher in Another Age of Lost Media

There are all sorts of reasons to be against the merger of Paramount and Warner Brothers-Discovery. There's the obvious problem of increasing concentration in the already pretty-concentrated film and TV industry. People have problems with Paramount's owner Skydance interfering  with its news coverage and late night comedy. But one of the biggest is the way we are losing choice and access to information and culture.  

In the context of the digital world, there are two major problems with these mega mergers. One, as we've said in the past, is that it further degrades the already awful state of mainstream streaming services. In short, by increasing the amount of legacy content under their control, major studios don't have to invest in either new content or well-engineered services to get users. Plus, they can pool the data they have on users together in an increasing assault on consumer privacy. A monthly subscription and more information about your habits and interests? Win-win for them.  

Another problem is the disappearing of content. In the digital era, there should be no more lost media. Making and storing copies of media is easier and cheaper than it ever has been. The idea of the persistence of content online has become the adage "the internet never forgets."  

However, since more and more media is locked behind subscription services under the control of fewer and fewer companies—and as physical copies, or even just digital copies that are actually owned by consumers also vanish—access to our common culture is at risk. Incentives to take information and media offline should raise huge red flags. Unfortunately, they are multiplying. 

First, there's the impulse of merged companies to offset the cost of their merger by shelving projects and claiming them as tax write-offs. We saw this the last time Warner Brothers was taken over, with projects being ended in order to take advantage of a specific kind of tax benefit. It has little to do with the quality or even how much or how complete a project is—money is the driving factor and we as a culture are all poorer for it. And we all know it. That's why the outrage eventually led to Coyote vs Acme being released by a much smaller studio. But not every project can count on having the perfect storm that brings a movie or show back from the dead.  

Second, there's the imposition of a new owner's own goals and beliefs upon its new acquisition. Some choices might seem kind of benign, like Discovery axing a lot of family-friendly and kid-oriented content from HBO Max. Of course, when one of the things affected is a pillar of childhood like Sesame Street, the loss of access to old episodes becomes a loss to the culture and not just a change in direction.  

 Another area where we might lose access to information is in the radical shift in direction for news reporting. Reporters have resigned and more are being offered buy-outs to support the change in direction. Whether this will escalate to older projects being removed from streaming services or quietly buried in the recommendation algorithm is yet to be seen, but remains a real possibility. And it's not lost on anyone that one of the outlets reporting on the chaos at CBS News was CNN, which is now owned by the same people who own CBS News. The loss of diverse sources of news, as imperfect as we may have thought they were, hurts all of us. It makes it harder to access information and diverse points of view.  

Finally, media disappears as licenses expire and as studios decide they no longer want to pay residuals or backend deals to the creative people who made the work in the first place. Again, this looks like business decisions but if the works were exclusively available via streaming, those movies disappear forever. That both contributes to the idea that works are disposable and that paying the workers who made the art is not as important as shareholder value. 

On top of all of this, draconian copyright regimes and the anti-user product designs they create have made it nearly impossible for people to archive this material on their own—as many used to using VCRs. It’s easy to see a crisis in access to culture and information brewing, at the same time that it should be easier than ever to preserve and make available our shared digital culture.  

Just do it for me? Balancing human and AI input in content improvement

Working with student interns and contributing to the Drupal community, I’ve experimented with AI in various  tools designed to support content improvement. I reflect on what I’ve learned about peoples’ preferences for engaging with AI tools and consider wider design challenges for web content to support co-browsing.

Writing good quality, user-centred content has always been a challenge, and accordingly, a suite of tools evolved to address this; Hemingway, Grammarly and many more. With their ability to learn patterns and predict word sequences, the advent of LLMs provided a new direction for content improvement. Where previous tools had used fixed rules, algorithms and formulas to flag content problems and score sentences, LLMs changed the playing field as they were able to be prompted to rewrite and create text and to adapt words and sentences not only in line with fixed rules, but also aligned to different interpretations of style.

The UX Service’s content design course, Effective Digital Content, and our ongoing Content Improvement Club meet-ups are both designed to help staff publishing digital content for the University. It’s been encouraging to see a positive uptake for both training formats, however, we realise we can do more to help staff publish the best quality content they can, and to this end, for the past few years it’s been a privilege to support a series of interns in the UX Service to look at how we can innovate and apply AI to the persistent challenge of writing good quality, user-centred content.

Thinking beyond – embracing the spirit (or grasping the nettle, depending on how you look at it) of AI innovation, it’s sensible to note that, more broadly, the browsing landscape of the web is increasingly shifting to include AI agents as well as humans, giving cause to consider what using AI to design content to be consumed by AI might look like. That’s the subject of another blog post (or several).

For now, as I begin another phase of work with a new student in the role of AI and UX Innovation intern (Viktor Sharha), it felt timely to pull together what I’ve learned so far.

Designing, creating and editing content happens in various ways

When designing any user-centred assistive features, a good place to start is the current workflow or user journey people adopt to achieve the task in question. In other words, to look at the stages content editors currently go through to make sure their content is as good as it possibly can be.

One of my earliest research tasks when I joined the University was to usability test the Editorial Style Guide, to find out how publishers used it to keep their content consistent in EdWeb. This work gave me insight into the type of workflows publishers followed to get compliant content published. Some publishers worked in Microsoft Word, using the in-built spelling checker for typos, and cross-checking against a printed-out copy of the style guide for University-specific errors before pasting content into the EdWeb body text. Others went straight to EdWeb, checking their content line-by-line with a PDF version of the guide. Others relied on colleagues giving their work the once-over before pressing publish. Some used Grammarly and Hemingway, some didn’t.

Overall, there wasn’t a dominant content-publishing workflow, with unified friction points waiting to be alleviated through a standard user-centred design process. Instead, there were varied approaches, habits and ways of working which people had adapted over time to suit their specific circumstances and to address the areas they found difficult as individuals. Tools were used at different stages of the process, and to a greater or lesser extent depending on personal preferences.

Therefore, designing content improvement tools isn’t one-size-fits-all

It follows that, since people follow different pathways to create, design and edit content, tools to assist this process need to be developed in different ways to suit varied preferences. Since 2025, interns in the UX team have prototyped our own different content improvement tools to test with publishers, and we’ve also tested open-source tools that exist already, developed within the Drupal community. From our testing and research, we found pros, cons and nuances associated with different tools. The summary below pulls together detail of the tools and how they compared.

Tool 1: Drupal Editoria11y – no AI, flags errors in the editing interface

Editoria11y is a Drupal module built by John Jameson from Princeton University that automatically checks content inside the editorial interface. It was originally designed to catch accessibility issues, but it can be configured to check content against other rules too, such as institutional style guide conventions.

What it does and how it works: Editoria11y checks content as authors type, and also on previews and published pages. It includes an indicator bar to show how many issues are on the page and has inline markers to show where each issue is, as well as a dashboard summarising site-wide errors. When an issue is opened, a pop-up modal explains the problem and suggests a fix.

What worked well: In tests, publishers were able to make sense of the inline markers and understand how these related to the indicator bar for the whole page. They liked the tracked change visualisation in the pop-up modals, and were able to work through successive modals, accepting or ignoring the suggested changes.

What needed improvement:  Some publishers found the number of markers overwhelming, and since the pop-ups covered the text, it was difficult to judge each recommendation in context. In the pop-ups, some of the rule descriptions used formal grammatical terms (such as ‘ordinal suffixes’) which were unfamiliar to some publishers meaning they were confused about the problem to be fixed. Since rules were applied mechanically, there were instances where errors were missed, or where awkward fixes were suggested (such as swapping ‘1st floor’ for ‘1 floor’ due to straight removal of the ordinal suffix) – these degraded publishers’ trust in the tool.

Read more about what we found testing Drupal Editoria11y

Embedding the rules: What we learned UX testing a style guide helper tool built with Drupal Editori11y

Tool 2: ELM-powered content design chatbot – with conversational advice

This tool was developed by student Mostafa Ebid during a 2025 summer internship. It connected ELM, the University’s AI platform, to EdWeb2, so publishers could get help with their content by conversing with the University’s suite of LLMs without leaving the Drupal editing interface.

What it does and how it works: This tool sits in a chat sidebar next to the main Drupal editorial interface and offers three modes: Write Content, Design Content and Proofread. In response to a user prompt to improve content, it activates an orchestration of agents (each with specific tasks such as checking the style guide, structuring pages, profiling audiences) which read the page content and act on it accordingly, publishing the revised content in the chatbot window.

What worked well: In tests, publishers understood how the tool could be used as a writing partner, and were impressed with the way it restructured content with headings, bullet points and calls to action and hyperlinks.

What needed improvement: When faced with the empty chatbot interface, some publishers weren’t sure where to start, and they didn’t notice the three modes to work on content. Some were also unsure of the sources the chatbot was using for its revised content and skeptical that the revised content contained hallucinogenic content. They would have appreciated tracked changes to be able to pinpoint what had changed and found the long thin sidebar chat interface awkward to review the updated text. The chatbot was also limited to reviewing body text content, whereas publishers needed it to review content in other page components used in Drupal.

Read more about what we found developing and testing the ELM-powered content design chatbot:

Initial insights from UX testing our Drupal AI content assistant tool

Tool 3: QuillMark v1 and v2 – some AI, clusters errors with correction suggestions

QuillMark was an AI-enhanced tool which worked in Drupal to check content against the University’s Editorial Style Guide. Shlok Gupta built and tested two iterations of Quillmark during his 2026 summer placement as AI and UX Innovation Intern.

What it does and how it works: QuillMark sits in a side panel of the Drupal editorial interface and scans content on the page in two stages. The first stage uses code-based pattern matching (without AI) to catch straightforward errors. The second stage brings in AI to find deeper issues (such as passive instead of active voice) and to suggest fixes. Findings from both stages are displayed in the side panel – each finding describes the problem, links to the relevant rule in the style guide and offers an ‘Apply fix’ button (which can be reverted). Findings are grouped by category, and this panel also includes an indicator of the environmental impact and cost of the scan.

What worked well: In tests, publishers understood that the tool was there to help them correct content on the page, they liked the environmental and cost indicator, the suggested fixes and the links to the style guide reassured them that the tool was applying University-specific rules.

What needed improvement: Publishers reported feeling overwhelmed by the number of findings on a single page and they wanted a way to deal with related findings and fixes in bulk, to avoid going through them one-by-one. Related to this, they wanted

Read more about what we found developing and testing the QuillMark tool:

QuillMark, iterated. How user research helped us evolve an AI-enhanced Drupal style guide assistant

Tool 4: Drupal AI Content Review – full AI, with a plan provided to improve content

As part of the Drupal AI Initiative, I’ve contributed to shaping the self-explanatory Drupal AI Content Review module. This module plugs into a Drupal site, references content rules and good practice from context items in Drupal Context Control Center (within the same site) and uses the configured AI provider to review site content

What it does and how it works: The feature reads the page content and checks it against criteria set by the site owner, such as tone and voice guidelines, inclusive language rules and target readability levels. These rules are stored in the Drupal Context Control Center, along with examples of good and bad content. The review gives the page a score for each criterion and recommends fixes to improve it. Accepted recommendations become an improvement plan, which is applied in the editorial interface. Re-reviewing the page shows the updated scores, with the option to keep making improvements. Over time, the rules can be refined by adding more examples of good and bad content to the Context Control Center.

Read more about Drupal AI Content Review on Drupal.org

Read more about the Drupal Context Control Center on Drupal.org

View a demonstration of AI Content Review in a mocked-up university website as part of the Drupal AI demo on the Drupal Forge platform

What worked well:  In tests of this module within the AI demo, people liked seeing the red, amber and green indicators against page content for an easy visual highlighting areas to be worked on, and they liked being able to select a checklist of recommended improvements to apply to improve their content.

What needs improvement: The screens explaining the recommended improvements were difficult to scan read, and it was also difficult to compare the original text and the text with the recommended improvements side-by-side (despite colour shading to denote which was which). The stage where the recommendations were turned into an improvement plan felt like an unnecessary extra step, it would have been better to have implemented the recommendations directly. Also, it was presumed that implementing the plan would have corrected all errors resulting in a score of 100%, without the need for further work and re-review. Finally, it wasn’t immediately clear how to use the Drupal Context Control Center and the AI Content Review together.

When designing AI-enhanced tools, consider three AI design maxims

Taking on board variations in blending human effort and automated operations, there are three approaches to designing AI features and AI-assisted workflows worth considering. These are defined below, with ideas for how they may apply to content improvement.

1.      Human-in-the-loop

In this set-up, humans typically share the task-load with AI – actively working alongside an automated operator to supervise or influence actions, and to approve AI decisions before they happen (ensuring AI outputs are not released without the green-light from a person).

Human-in-the-loop tends to be adopted when:

  • there is a need for human judgement to get something correct
  • the consequences of an error could be high-profile or damaging
  • trust in an automated tool is still being built, and the boundaries of its capabilities are still being realised
  • the goal is for the system to facilitate shared learning between the person and the automation

For content improvement, human-in-the-loop is appropriate in these scenarios:

  • ensuring correctness of important factual data – for example, prices, dates, contact details, terms and conditions
  • checking the tone of sentences in line with inclusive language guidance
  • verifying accuracy of high-stakes published content such as legal information, policy wording or written instructional content
  • rewriting complex topics in plainer language
  • writing meta descriptions or page summaries to be accessed by search mechanisms
  • translating content into other languages

2.      Human-on-the-loop

These systems involve the AI or automator acting on its own, with the human reviewing and spot-checking at intervals, monitoring output quality over time, and intervening to correct failures periodically, where required.

Human-on-the-loop tends to be adopted when:

  • there is a high volume of low-risk tasks to be completed, which would be tedious for humans to do themselves
  • consequences of errors are minimal
  • there’s a need for tasks to be completed quickly, and including human-input at each task would be too time-consuming
  • consistency is required across tasks, and there’s a risk of human error introducing variability

For content improvement, human-on-the-loop is appropriate in these scenarios:

  • fixing straightforward style rule violations such as correcting spelling errors
  • categorising and tagging text content (for example, to facilitate its discovery in search)
  • scanning content for readability and accessibility, to flag or report areas to be addressed

3.      Human-out-of-the-loop

In this arrangement, the human sets up the AI to act autonomously, and leaves it to proceed, with no routine review.

Human-out-of-the-loop tends to be adopted when:

  • there are very low-risk, behind-the-scenes tasks to be completed where errors have little or no consequence and can be fully reversed

For content improvement, human-out-of-the-loop is appropriate in these scenarios:

  • autocorrecting typing errors
  • detecting content duplication
  • auto-optimising media content to defined parameters

Choice, pacing and transparency are constant design requirements

Testing four different tools has showed that options always need to be available, whatever balance of human and AI the tool includes. People approach AI differently based on preferences and past experiences. Some love nothing more than a good grappling of a sticky sentence, some just want an agent to do the work for them, and many sit somewhere in between, happy to have some AI help as long as they have the final say.

Alongside humans in, on and out of the loop, it’s important to remember the fourth option: human only, where people can choose to do their content unaided, using the methods they relied on before AI came along. AI shouldn’t be forced on anyone, so all tools need to come with a ‘no thanks to AI’ option.

AI is a fast-evolving and non-deterministic technology, so at any point in time it may work differently to the way it did before.  Responding to challenge from a design perspective, it’s important to factor in pacing. People need to be taken through how a tool works, and the options available. They need time and space to experiment and try it out, so they can learn for themselves how it can work for them. Pacing not only matters in the way tools are introduced, but also in the way their outputs are presented, to avoid the overwhelm that comes when AI does the heavy lifting all at once, and so people know what changed and how, and they have a way to accept, decline or edit what it produced.

If people don’t trust an AI tool, they won’t use it. Earning trust requires transparency, explainability, relatability and reliability – so AI tools need to be upfront about what they can and can’t do, with clear, intuitive interfaces and labels. Reliability will always be hard to promise given AI is probabilistic, so instead of accuracy guarantees which may fall short, a better approach is to design for experimentation, so people can find their feet and make their own judgements.

Designing for humans and AI together – metadata, markdown and MCPs

Looking at the broader digital landscape, the challenge widens from designing tools that help humans make content to designing websites with content ready for both humans and agents to use.

In co-browsing, people are increasingly arriving at a website with an agent alongside them, handing off some tasks while keeping control of others and wanting to see what their agent is doing. That means firstly deciding which tasks belong to humans and which one to AI, and secondly designing for each. Humans need intuitive interfaces. Agents need data they can read and act on without ambiguity:

  • metadata, such as schema markup that tells an agent what a page is about, who it’s for and how it connects to other content
  • markdown versions of pages and llms.txt files, which give agents clean, structured text without layout and navigation getting in the way
  • MCPs (Model Context Protocol servers), which let sites expose what they can do as well as what they say, from simple actions declared directly in HTML to more complex ones written in JavaScript – in preparation for use by the emerging WebMCP browser standard.

Labels matter for humans and agents alike, because words are becoming a currency that machines read as well as people. Alignment on terminology facilitates the creation of a single source of truth, to prevent the human experience from drifting from the agentic experience. On another level, humans need a way to consistently communicate with AI – to give AI the rules and instruct it and to train it with new insight. ‘Rosetta stone’ mechanisms like the Drupal Context Control Center will play an important part in designing for human and AI interaction and collaboration and I am excited to keep working and researching in this area to address these evolving design challenges.

 

Show me your sources! What students make of content provenance data, AI signifiers and sparkles

The line between human-made and AI-generated content is blurring fast. I added a disclosure indicator to an AI-powered search tool and tested it with students, which taught me about perceptions of AI-influenced content and what signals authenticity. I reflected on the need for a shared language around provenance and disclosure.

Estimations of the proportion of AI-generated content on the web vary and are becoming progressively difficult to measure. New legislation advocates for disclosure of AI use, but the nuances of communicating how AI has been used within a feature, and detailing the finer points of the provenance of AI-generated content both remain a challenge.

Earlier in 2026, I took part in a study to design a C2PA provenance indicator into a prototype AI searchbot. I revisited my designs to test them with students, to assess if this authenticity data made sense and to understand its impact on students’ potential trust of the AI searchbot. In this blog post, I describe what I tested and what I learned, and make broader considerations of designing with provenance data, the impact of AI legislation and what makes people trust AI features.

People’s trust in technology depends on explicit and implicit factors

The ISO 9241-210 standard for UX defines user experiences as ‘a person’s perceptions and responses that result from the use of anticipated use of a product, system or service’, which sums up the influence trust can have on the way a person experiences a technology.

I was interested to explore more about the relationships between human trust and AI, so I read a selection of papers on the topic and extrapolated four factors that commonly affect how people trust AI technologies:

  1. Transparency (surface and presumed credibility)
  2. Reliability (earned and reputed credibility, how consistently it works and how dependable it is)
  3. Explainability (how clear the purpose of the AI is and how clear it is how it works)
  4. Relatability (how well the AI supports the person’s interests, how easily they can use it and how it resonates with them)

When it comes to designing AI features people trust, and communicating how AI has been used within those features, it’s important to try and convey each of these factors in ways that make sense to the user in context. In other words, striking a delicate balance of providing sufficient trust indicators to allow people to make reasoned judgements, without going overboard and running the risk of confusion or skepticism.

The EU AI Act has set out obligations to disclose AI usage

From an ethical perspective, advising people that they’re using an AI feature has always been the right thing to do, in order to foster a sense of user trust. Legislation published in August 2026 made the requirement to disclose AI usage more explicit and provided more detail. Article 50 of the EU AI Act details Transparency Obligations for Providers and Deployers of Certain AI systems and includes recitals which cover areas including the following:

  • Designing AI systems to ensure users know they are talking to AI
  • Marking content from AI systems (text, images, videos, audio) so it is detectable as such
  • Informing people of AI systems that include emotion recognition or biometric categorisation
  • Disclosing deepfakes (audio/video/image content that has been manipulated as well as public-interest text, unless human-reviewed with editorial insight) as AI-made
  • Clarity, timing and accessibility of AI disclosures – to be provided at the first interaction or exposure and to be clear and accessible.

With this legislation now in place, it’s anticipated that a growing number of ways to disclose AI usage will emerge, making AI disclosure as mainstream as AI features.

I adopted the C2PA digital content provenance standard into an AI searchbot prototype

When it comes to icons that disclose AI usage, sparkles, neurons, and robots are all widely used, but they only signify that AI is present or available, they don’t provide any detail of how it has been used or how it has shaped content. C2PA is an established provenance standard which has been designed to convey progressive levels of content provenance data. I incorporated C2PA to disclose how AI was used to shape synthetic chatbot responses in a search feature and wrote about my process in an earlier blog post:

How can people trust AI-generated content? Designing provenance data into our prototype AI searchbot

I concluded this design work with a resolve to test my designs with users to understand the following:

  • if they cared about seeing provenance data about the AI searchbot response
  • if they grasped the concept of provenance data
  • if they understood what the C2PA Content Credentials icon signified
  • if the names for the different categories of data made sense to them

Drawing on my recent reading into trust factors, and taking into account the new EU AI disclosure guidance I also wanted to use my designs to understand more about how users make sense of AI features, and what affects their trust of them in context.

Testing revealed students’ responses to progressive levels of AI disclosure in the searchbot

In a simple test, I asked the student participants to imagine they were in a scenario where they had used the AI searchbot to find information from a website. I then showed them mock-up screens with the indicator in progressive levels of disclosure, going from the C2PA ‘CR’ logo, to ‘content credentials’, ‘About this response’, ‘Certification data’, ‘Digital source data’ and ‘Transformation processing data’. At each stage, I asked them to speak aloud to describe what they saw, what they thought it meant, and how it resonated with them.

Screenshot showing the AI searchbot interface showing all the provenance data available

Screenshot showing the AI searchbot interface with progressive levels of provenance data on display

The C2PA icon was unfamiliar but content credentials made more sense

Students’ first exposure to the provenance data was as the C2PA ‘cr’ icon which was displayed at the opening interface of the searchbot (when the searchbot was ready to receive a user query). On its own, the icon was pretty meaningless to them. One felt it could be a watermark, others thought it was a button to make the chat window bigger, others said they would Google ‘cr’ to find out what it was. The next level of disclosure was a signifier that read ‘content credentials’, which appeared once the searchbot had received a query and produced a response. Students found this indicator more meaningful – they felt it was a way to verify where the response came from, and said they would be inclined to press it to find out more.

The level 1 disclosure ‘About this response’ was too vague to be useful

When asked what they thought they would find in the section ‘About this response’, students were largely unsure. Some thought it would contain a hallucination warning, others thought it would contain sources. When they read the generic descriptor about the C2PA provenance data, which explains that the response was AI generated and had content credentials and assertions to demonstrate authencity, they felt it was too vague to be useful to them, and didn’t feel it was telling them anything new. They were unfamiliar with ‘content credentials’ as a term and by the language, they reasoned that this message was aimed at a legal team rather than someone reading the output of the searchbot.

‘Certification data’ wasn’t a term that resonated with student participants

Reading the label ‘certification data’ students had mixed ideas of what this section would contain. Some felt it would reveal intellectual property details, some thought it would show licensing agreements relating to data entered in the searchbot, and others thought it would explain how the searchbot handled data. There was a further expectation that this section would contain the sources behind the searchbot response.

When they expanded this section to see the identity data (name of the organisation certifying the content) and the associated timestamp of when the content was certified, they were unclear what was being certified (the output of the searchbot or the AI model), and how they would use this information. Some reasoned that the responses from the AI searchbot were stored in a database and that this certification data could help keep track of a bank of these responses.

‘Digital source data’ made more sense to them

Of all the data categories displayed, ‘Digital source data’ was the one students gravitated towards. This is where they assumed they would find the links to the sources used to create the AI searchbot response, and their assumptions were correct. For many students, seeing the links to the sources proved the authenticity of the AI content more strongly than any other sections of the data they had seen. As well as the source links, this section included the C2PA Actions assertion (which recorded the stages used to create the digital content (or asset) including the AI model used). Students recognised the AI model but questioned the terminology ‘Actions assertion’, thinking it would be more straightforward to just list the AI model used.

In the ‘Transformation processing data’, detail of the AI prompts was of particular interest

Preconceptions of what they would find in the ‘Transformation processing data’ section varied from a logical set of steps to show how the responses were formed from source data, to details of how data inputs would be used to train AI models, to sustainability data showing energy expenditure in the formation of the AI searchbot responses. When they saw information labelled ‘Search mechanisms’ which detailed how embeddings were created and a vector database was used in semantic matching, most students felt this was too much technical detail. Some who were familiar with the process felt this section would be more appropriately labelled as a technical stack, rather than search mechanisms. The ‘Transformation processing data’ section also contained the AI prompt, and most students responded positively to this, saying they felt it gave them a fuller picture, it helped them understand ‘the personification of the AI’ and made the process more transparent.

The effectiveness of my C2PA designs largely depended on students’ existing knowledge

Reviewing the findings together, I concluded that my designs partly succeeded in communicating details of how AI had formed the searchbot responses, but this depended on students’ perceptions and what they knew already, as well as what they wanted to know. Overall, my designs fell short when it came to instilling a sense of trust in the students, largely due to unfamiliarity with the C2PA terminology, the challenges of targeting implicit trust factors and the partic.

Students familiar with AI responded positively, but others wanted a non-AI option

The C2PA indicators focused on the provenance of the AI-generated searchbot responses, but for some students this was irrelevant as they wanted a choice to be able to not use the AI feature at all. Several students were skeptical and negatively pre-disposed to AI features based on their previous experiences and concerns about the environmental cost, and therefore upon seeing the sparkle icon realising the searchbot was AI-powered they were sought for a way to turn it off, or to make a choice about whether to use it or not. This reinforced the need to ensure AI features are a choice, and in the case of the AI searchbot example, an alternative way to achieve the same task is provided.

The C2PA terminology was off-putting to students

As they read the data in the different categories, students felt alienated by some of the C2PA language, particularly terms like ‘content credentials’, ‘provenance’ and ‘assertion’. When they were able to see contextual detail associated with these terms, most were able to fathom what the terms meant, but I concluded that including these terms in the designs and in the labels of the different data categories was giving students an extra unnecessary hurdle of jargon to overcome before they could reach an understanding of the data presented to them.

Some C2PA data was clear and useful, other data felt unnecessary

The four categories of data resonated differently with the students, and, based on their reactions, some of these data sources felt unnecessary to show data provenance.

Most students found the source links useful, and what they expected in the category labelled ‘Digital source data’. They also found the AI prompt useful, but they didn’t expect to find it in ‘Transformation processing data’. Almost all felt the notification that the response was AI-generated (in the ‘About this response’ section) was unnecessary. Other data points that students felt confused and nonplussed by included the ‘CR’ logo, the identity data, the certification data (including the timestamp), and the technical terms describing the response-generation process.

Some of the data students were expecting was missing

As they reviewed the different categories of data they were presented with, students commented on data they would have expected to see. This included: environmental impact data, information about how input data would be used to train AI models, and details of how the data was processed and held by the searchbot feature. While these data points largely fell out of the realm of content credentials, it was interesting to understand the types of data students looked for when making judgements about whether to use and trust this type of AI feature.

The C2PA designs targeted some trust factors but not others

Re-examining factors that affect trust highlighted that trust is heavily influenced by implicit factors or emotional cues which are specific to individuals. Targeting subjective trust factors such as relatability, (which depends on how well something supports a person’s interests and resonates with them) was very difficult to achieve without research to understand the nuances of the target audience. Some students were reassured by the mention of AI models involved in generating the content, but for others, this merely caused confusion.

Targeting explicit trust factors like transparency and explainability in displaying content provenance was more straightforward, but it meant striking a balance. Providing too much technical detail about the production process, although this was intended to build trust by being upfront, could end up causing confusion and therefore result in the opposite effect, as I observed in the student tests.

Conveying reliability remains a challenge when AI is involved. Since AI is probabilistic, its results are variable, and in the context of a feature like an AI searchbot, it was not possible to make assurances in the provenance data. Instead, I used the progressive disclosure levels to encourage exploration and a gradual building of understanding, to help people come to their own conclusion about whether to trust the AI-generated content. I saw this working in the student tests, as they made reliability judgements about the AI responses when they saw the source data revealed.

Provenance indicators should replace the sparkles and it’s time for a unified AI vocabulary

The C2PA standard is several years old, yet content provenance is still a relatively unfamiliar concept. With new AI legislation published and the advent of more formalised requirements to disclose AI usage and application, I predict more rhetoric and discussions around disclosure, authenticity and provenance, to evolve the language and terminology around these concepts and to make them more mainstream.

Those who consume AI-generated content want to know how it was made, and the sources it came from, and those who use AI to generate or shape the content they publish have an obligation to disclose how AI was used, to ensure transparency from both perspectives. When designing AI features, consistent labelling will play an growing part in conveying meaning. As people become more familiar with AI, they will rely on these labels more and more to make judgements and decisions about AI features and AI-influenced content.

This small study has re-emphasised the importance of choosing terms and wording labels carefully, to strike a balance between technical and non-technical and to provide just enough detail to resonate with audiences but not so much that it confuses them. It highlights the need for a unified vocabulary for AI, to effectively communicate about its usage and application. As AI usage progresses, I expect provenance indicators and associated design patterns to continue progressing, and I’m keen to keep a close eye, to keep iterating on ways to display provenance data. I look forward to seeing sparkles replaced by more meaningful signifiers of AI usage and disclosure.

Google Summaries Aren't About AI, They're About Lock-In

Whether they call it AI summaries, knowledge panels, or AMP, Google takes actions that are less about functionality and more about locking you into its own ecosystem. Knowing that makes it possible for all of us to choose better services than Google's, defeating their anticompetitive goal. In this case, as it is so often, knowledge is power.  

In the old days, when dinosaurs roamed the earth and getting online involved your computer making eldritch shrieking, Google was a website that took whatever you typed in and gave you a list of links that best matched what you were searching for. As it's grown beyond that, however, the core function of Google Search has moved from giving you the best answer to your query to a number of other things, most of them about making money for Google. 

This is obviously not a new observation. It's at the core of what EFF Special Advisor Cory Doctorow named "enshittification." But it's important to remember that whatever shiny new name or technology is wrapped around these tools, it's not about the wrapping, it's about the monetization. 

In the case of Google Search, the company has a history of rolling out new "features" that are less helpful to the searcher and more designed to keep you and your data within the Google ecosystem. One easy example of enshittification is putting sponsored links and ads at the top of search results, forcing you to scroll down past a bunch of not relevant links to get to your answers. Or, in some cases, accidentally clicking on sponsored links that might be designed to give you the opposite information of what you're looking for or for a competitor to the product you intended to buy.  

The other clumsy version of this is Google's self-preferencing, where Google's own services—say YouTube if you're looking for a video—are weighted so that they appear at the top of a search over a competitor's. Similar to this was Google's push for websites, particularly news ones, to be made available via what it called "Accelerated Mobile Pages" or AMP, which in theory stripped out features that might make loading a page on mobile too slow, but also handed over control of how webpages looked on mobile to Google and turned mobile links, often by default, into Google links.  

But the other way that Google has historically kept your—and your valuable eyeballs—on its website has been to just obviate the need to even click on a link and leave Google at all. The first version of this was what Google calls "knowledge panels." These are snippets of information shown on the results page of a search that purport to answer your question. Knowledge panels predate the AI push and haven't always been what you could call accurate.  

Remember that it is Google's monetary incentives that drive the decisions and not necessarily the tech itself.

Knowledge panels are extremely similar to the AI summaries we now see at the top of Google Search results, except that instead of pulling quotes from specific linked sources, it's the AI doing the summarizing, though also sometimes with links. They are basically the same thing, mostly changed only in branding, because AI is now the cool kid on the block. The goal is essentially the same—it's less about providing you with the best answer to your query and more about making you think it's unnecessary to click a link and leave Google's ecosystem. And, more importantly, it keeps you on Google's pages longer to read the content. More time on the page is more time exposed to ads and sponsored links and more money for Google.  

The biggest irony here is that it didn’t even work. People choosing to use “AI mode”—and possibly made accustomed to getting information as text instead of links through the knowledge panels and the AI summaries at the top—has caused “lower monetization” via ads. Ads, remember, are Google’s main source of revenue. Like newspapers before them, Google is finding that counting on a monopoly on ad space maybe isn’t the most sound strategy in a shifting technological landscape. And to get out of the hole it has made for itself, it is lighting cash on fire to try to get back on top. Perhaps they should consider pivoting to video.  

Every time you see a feature like this, no matter what it is called or what "new" technology they claim fuels it, remember that it is Google's monetary incentives that drive the decisions and not necessarily the tech itself. And you can always choose to ignore it, scroll past, disable it, or even go to another service entirely.  

QuillMark, iterated. How user research helped us evolve an AI-enhanced Drupal style guide assistant

AI and UX Innovation Intern Shlok Gupta and I worked together to build and test Quillmark, an AI-powered Drupal content design tool that tackles the persistent problem of keeping content consistent with the University’s style guide. This post covers how we iterated from version 1 to 2.

Leading a UX Service, it’s critical to stay informed about the latest technologies, as these dictate user expectations and perceptions. I recognised the potential for AI to transform publisher experiences of using Drupal when Drupal AI experts Freely Give came to work with our team in January 2025. Since then, I’ve been committed to experimenting with AI application in content management, both through my involvement with the Drupal AI initiative and by supporting a succession of interns in my UX team at the University.

As a technology, AI is becoming increasingly mainstream. That said, its rapid rate of evolution and flux means adopting it consistently remains a challenge, as does designing AI features that are centred around user needs and expectations. AI features can be built and released in a matter of days, but can just as quickly become superseded and obsolete as technologies and capabilities change, and user expectations and perceptions of AI features can vary widely depending on individual experiences, trust factors and preferences. Continuous research is necessary when developing any user-centred software, but it’s particularly important when designing user-centred AI features, which I wrote about in a related blog post:

Stop chasing, keep researching: Why continuous contextual learning is the only way to build useful AI features

In the spirit of continuous research, the students who occupy the AI and UX Innovation Intern role in my UX team spend as much time testing as they do building. Shlok Gupta joined the team from May – August 2026, in which time he prototyped and tested two versions of an AI-enhanced style guide checker tool in Drupal. Shlok described his process to build and test Quillmark v1 in two blog posts, and also produced a video to show it in action:

Building QuillMark: testing a Drupal style-guide assistant with web publishers

Quillmark: the technical details behind my Drupal style-guide assistant

View a video of Quillmark v1 on MediaHopper

In this post, I continue the story, to explain how version 2 was built based on feedback from version 1, and what we learned testing version 2 with publishers.

QuillMark v1 had a non-AI stage and an AI stage and the ability to flag and fix errors

Shlok designed QuillMark v1 to have two stages as a way of making most economical use of AI time and tokens. Stage 1 used basic pattern-matching and code-based rules to catch straightforward style-guide errors without AI, and stage 2 then brought in AI to find and recommend fixes for deeper issues. This meant AI resource wasn’t spent identifying and resolving simpler problems that didn’t need it. The screenshots indicate how v1 worked.

Screenshot showing Quillmark v1 (in the right-hand panel of a Drupal site) ready to review content

Screenshot showing Quillmark v1 (in the right-hand panel of a Drupal site) ready to review content

From right-left, four screenshots showing detail of the QuillMark guide in the right-hand panel before a scan, the panel expanded to show the guide, results of a stage 1 scan and detail of selected stage 1 findings

From left to right, detail of QuillMark in the right-hand panel – before a scan, expanded to show the guide, results of a stage 1 scan and detail of selected stage 1 findings

Three screenshots of QuillMark, showing the panel ahead of running the stage 2 scan, combined results of the stage 1 and stage 2 scans and detail of selected stage 1 and stage 2 findings alongside each other

From left to right: the panel ahead of running the stage 2 scan, combined results of the stage 1 and stage 2 scans and detail of some stage 1 and stage 2 findings alongside each other

QuillMark screenshot showing how a finding from a stage 2 scan mapped to a sentence in the passive voice highlighted in the editorial interface

QuillMark screenshot showing how a finding from a stage 2 scan mapped to a sentence in the passive voice highlighted in the editorial interface

QuillMark screenshot showing the highlighted sentence changed to active voice - which resulted from pressing the ‘Apply fix’ button on the finding

QuillMark screenshot showing the highlighted sentence changed to active voice – which resulted from pressing the ‘Apply fix’ button on the finding

Tests revealed the need to simplify QuillMark v1

In the tests of QuillMark v1, participants were asked to read a guide about how the tool worked, use it to scan some pre-prepared content and then review the resulting findings and say how they would act upon them. The results indicated several areas for improvement.

All participants understood that QuillMark v1 was a tool to help them correct content errors. Despite being directed to read the guide upfront, however, some participants were unclear of the difference between stages 1 and 2.  Having run stage 1, participants were uncertain what stage 2 of the tool would do. Some identified the AI detail in the stage 2 labels and therefore realised that stage 2, unlike stage 1 used AI, but others thought stage 2 was the ‘repair’ or ‘fix’ stage to follow stage 1 as the diagnostic stage.

As they reviewed the different categories of finding in the right-hand panel, most participants said they felt the amount of information was overwhelming, and difficult to scan read. They especially found it difficult to interpret abstracted recommendations to fix errors, preferring to receive suggested fixes through the ‘Apply fix’ option. As a way to reduce the overwhelm, some participants said they would be inclined to go and fix several of the errors manually first, and then rescan, so that the tool only presented them with the errors they missed themselves.

Other observations from the test participants fell into two categories, lack of understanding of how Quillmark worked, and feedback on how they felt it could work better for them.

Participants were confused about how to use QuillMark and how it worked

Several aspects of participants’ behaviours indicated they weren’t sure how QuillMark worked. These were as follows:

  • There was some confusion over how to start QuillMark running. Participants did not immediately realise that the Stage 1 and Stage 2 buttons would start the tool. They looked for a button that said ‘Run’ or similar. When they worked out how to activate the different stages, they were unclear about what was happening in each.
  • Several participants raised questions about the cost and environmental impact of running the AI stage of the tool, saying this would affect whether they would choose to use it.
  • There was a lack of clarity from participants about whether QuillMark was using generic grammatical rules to flag errors and make recommendations or if it was using the specific University of Edinburgh style guide rules
  • Several participants did not initially recognise that QuillMark could suggest fixes as well as flagging errors.

Participants suggested ways to change QuillMark, to help it support publishing practices better

When they had got used to working with QuillMark, several participants suggested ways they felt it could work better. There were as follows:

  • A request to be able to apply bulk fixes to recurring errors, rather than addressing them one-by-one (for example, removing full stops in repeated headings or switching sets of times from the 24 hour clock to the 12 hour clock)
  • An ask from several participants to be able to preview fixes offered by the ‘Apply fix’ buttons before accepting them to change the text in the editorial interface (however, they soon realised they could switch between the ‘Apply fix’ and ‘Revert fix’ buttons to achieve this).

QuillMark v2 worked in the same way but the interface was simpler

Taking on board the feedback from the QuillMark v1 tests, Shlok made a version 2 which included several changes:

  • A single ‘Scan’ button to start both stages of the tool working, replacing the need for publishers to run stage 1 and stage 2 scans in sequence
  • Progress indicators that appeared once ‘Scan’ had been pressed, to communicate what was happening, and to affirm that the editorial style guide was being checked
  • Inclusion of an ‘Advanced settings’ panel giving users the option to decide whether to include deterministic error scanning and fixing and the use of AI to suggest replacements
  • A panel that estimated monetary cost and environmental impact of each scan resulting in the findings

View a short video demonstration of QuillMark v2 on Media Hopper

The screenshots indicate how v2 worked and show the differences between v1 and v2

QuillMark screenshot showing the tool in the right-hand panel ahead of a scan being run on a page of content titled ‘Research projects’

QuillMark screenshot showing the tool in the right-hand panel ahead of a scan being run on a page of content titled ‘Research projects’

Series of QuillMark screenshots, from left to right, showing the tool with the guide displayed, and the successive stages once ‘Scan’ has been pressed – reviewing, preparing, checking editorial style rules and then the notification that QuillMark has completed the scan.

QuillMark screenshots, from left to right, showing the tool with the guide displayed, and the successive stages once ‘Scan’ has been pressed – reviewing, preparing, checking editorial style rules and then the notification that QuillMark has completed the scan (showing that AI suggestions are included in the ‘Advanced settings’ pane

Tests showed v2 was better, but still had room for improvement

The testing protocol for QuillMark v2 remained the same as for v1, however, this time participants were asked to use the tool on three separate pages of University website content, to give them more practice of engaging with the tool to learn how it worked.

Participants found it easier to understand how QuillMark v2 worked

With the new blue ‘Scan’ button, participants were more easily able to get started with running QuillMark. Since it was running an AI and a non-AI stage QuillMark v2 took longer to produce findings than QuillMark v1, but the newly added progress indicators kept participants informed about what was happening.

The indicator with the data about environmental and monetary costs was noted by all participants, and some said this was welcome information to help them make responsible choices about AI use.

Participants were still directed to the guide, which some wanted to refer back to once they had got started with the tool, to remind them of how it worked.

There were fewer questions about the source of the rules powering Quillmark v2 than for v1, and several participants verified that the tool was using rules from the University editorial style guide by pressing the ‘Link to rule’ button on individual findings (which connected to the relevant page or section of the style guide).

Some still found the interface and display overwhelming

Although QuillMark’s interface had been simplified by the removal of stage 1 and 2 indicators, participants still commented it contained a lot of textual information and was difficult to scan. Some were unclear what the different labels in the findings meant (Errors vs Warnings vs Notes), and several found it easy to miss some of the tool’s functionality simply because there was too much to read through.

Participants were still uncertain about the ‘Apply fix’ button

As they worked through the individual findings, participants were able to recognise the need for corrections that had been highlighted, but some did not appreciate that QuillMark v2 offered a fix as well as flagging the error.

As they used QuillMark v2 to scan a page about research projects, the tool flagged an instance of poorly worded link text (‘Project Website’). The ‘Apply fix’ button contained a suggested improvement (‘Visit the Hillforts Atlas website’) but several participants did not select this, instead they manually addressed the error themselves, by checking the link destination and then writing appropriate link text. Those that did notice that ‘Apply fix’ had made a suggested improvement were happy to accept it and apply it.

Screenshot showing a QuillMark v2 finding flagging poorly written link text

Screenshot showing a QuillMark v2 finding flagging poorly written link text

Screenshot showing QuillMark v2 a fix to improve poorly written link text applied in the editorial interface

Screenshot showing QuillMark v2 a fix to improve poorly written link text applied in the editorial interface

They still wanted repeated findings to be consolidated, to be addressed together

Some categories of findings contained recurring instances of the same issue. Participants wanted these to be grouped or summarised instead of being listed individually, as they wanted to address all the findings on a particular issue together in one action (either dismissing them or applying suggested fixes), rather than going through them one-by-one. This problem was particularly prevalent on pages where the content followed a structural pattern, for example a series of headings all written in non-sentence case (with misuse of capital letters) or a labelling convention applied consistently to rows of a table (for example numerals in sequence).

Screenshot showing individual instances of an error flagged as separate findings in the QuillMark v2 interface

Screenshot showing individual instances of an error flagged as separate findings in the QuillMark v2 interface

There was concern over saving progress when working through lots of findings

In the tests, one of the pages scanned by Quillmark v2 generated 166 findings to be addressed. As participants worked through those in a particular category (for example ‘Links and calls to action’) they felt uncomfortable that they were unable to save as they went, seeking some reassurance that their work was being logged sequentially. They looked for a ‘Save’ button associated with each individual finding, and some wrongly assumed the ‘Dismiss’ button saved the change they had made (whereas in actual fact, this button removed the finding from the list to be addressed).

QuillMark was unable to reconcile multiple errors on the same content

In some instances, a piece of content contained multiple style guide errors, so it appeared in separate findings in separate categories. An example was ‘Prof Smith’s website’ used as a piece of link text. This error appeared in the ‘Spelling, wording and abbreviation’ category, flagged for the incorrect use of ‘Prof.’, but also in the ‘Links and calls to action’ category, flagged for unclear link text.

When participants encountered an error in one finding, they corrected it there, but then when the same piece of content was also flagged for another error in another category, Quillmark v2 had already noted that the content had been changed, and had no way of layering on another, different correction. In these instances, a warning flag disabled the ‘Go to field’ preventing participants from going to the incorrect content, therefore forcing them to try and find the location of the incorrect content themselves, which they found disorientating and tricky, especially in long pages of content.

Screenshot showing the blocker message that displayed when there were multiple errors on a single field of content

Screenshot showing the blocker message that displayed when there were multiple errors on a single field of content

Participants wanted QuillMark to fix heading levels and reposition inline links

Errors relating to content structure, such as incorrect heading hierarchies and inline links were flagged by Quillmark, but the tool was unable to suggest an automatic fix to apply. Several participants said they would have liked for an autocorrect functionality for these errors, since they found them difficult and fiddly to manage within the EdWeb2 editorial interface.

The next iteration of the tool will focus on one category of style rules: Links

Since the last round of tests on QuillMark v2, Shlok completed the internship, and was succeeded by Viktor Sharha in the role of AI and UX Innovation Intern, to continue the work on the Drupal style guide checker tool. Taking all the findings from the tests of QuillMark v1 and v2 together, I concluded that the most pressing problem to address was that of overwhelm.

When a page contains many errors or style guide violations, there is a dual design problem: Firstly how to detail each error, the root cause and the suggested fix, so that publishers understood what was wrong and how it can be corrected, and secondly to do this concisely, taking up the minimal space possible, to avoid publishers having a long scroll through multiple errors with the associated cognitive load.

Going forward, it made sense to focus on the first problem of presentation of individual findings first, and to do so at a site level (recognising that publishers may not do their checking on a page-by-page basis), before scaling this up to design for multiple errors. To maintain a tight remit, and to focus on the style guide violations which are highest priority to fix, we have decided to specifically look at links, to consider how the tool can spot instances of poor link text, how it can describe ways to improve these, and the options for applying automated fixes.

Site-Blocking Will Not Defend IP, No Matter the Bill’s Name

There has been a raft of site-blocking bills in the latest Congress, and the latest is called the “Deterring Extraterritorial Foreign Exploitation of Networks Damaging Intellectual Property” aka the “DEFEND IP Act.” The problem is that instead of “defending IP,” this bill will incentivize censorship, overblocking, and bad faith attempts to block access to a website. DEFEND IP Act, and all of these site-blocking proposals, threaten the open web. 

We keep seeing attempts to pass site-blocking legislation–from SOPA/PIPA in 2012 to Block BEARD, FADPA, and ACPA this year. Every one of them has at its core the rotten idea that enforcing copyrights requires building a censorship machine for websites into the architecture of the internet. This is, of course, a disaster for a free and open web. There is no way to create a mechanism for blocking access to an entire website that does not invite both deliberate abuse and lots of collateral harm to free and lawful speech.

DEFEND IP deputizes every service provider into a copyright cop, so long as a rightsholder has accused a website of copyright infringement. Let’s be clear: this isn’t about removing access to an infringing work–that already exists via the DMCA. This isn’t about getting damages from the website or the uploader. It is about making an entire website inaccessible for everyone trying to visit it.

DEFEND IP lets any rightsholder go to a court and get an order requiring service providers to block access to an entire website after alleging copyright infringement. What DEFEND IP does not have is any deterrent for someone seeking to block a website in bad faith. There are no punishments for getting a website blocked for protected speech. There are no meaningful remedies for those whose speech is vanished from the internet due to an entire website being disappeared. It creates a one-stop shop for getting an entire website–again, not an instance of infringement but an entire site hosting all sorts of user content–removed. But for those whose business, speech, or access to information is affected, there is no easy way to get the site restored.

DEFEND IP scales up the extraordinary legal structures that already exist for copyright enforcement. In doing so, it likewise scales up the problems those regimes pose to protected speech. 

We see this with DMCA takedowns all the time. We see it with bad faith takedowns used to silence criticism or commentary. We see it with the voluntary use of copyright filters by sites like YouTube, where seconds of sound matching seconds of sound in another video can prevent an entire work from reaching its audience. In these existing systems, there are at least some mechanisms of challenge available to the targeted creator. DEFEND IP has none. Instead, site owners, users, or readers will have to find a lawyer and go to court and hope to challenge the order, a slow, expensive, and daunting process

Those existing systems are already frustrating for the targeted creators and users, but under DEFEND IP a whole class of people doing protected speech will find themselves deplatformed because of the actions of others

This bill is not a defense of creativity or creators. It is a way to reshape the internet by building a vast new infrastructure of censorship. Congress should put aside DEFEND IP and the failed idea of site-blocking laws, for good. 

Congress Has Another Site-Blocking Bill, And This One Targets VPNs

Congress is taking another run at site-blocking, a deeply flawed concept that would undermine basic internet infrastructure. Rep. Darrell Issa (R-CA) has introduced the American Copyright Protection Act (ACPA), H.R. 10364, a bill that would give copyright owners a new legal tool to block Americans’ access to foreign websites accused of copyright infringement. 

The basic idea is all too familiar, and it’s still dangerous. A copyright owner first asks a court to label a foreign website a “foreign piracy site.” Once that happens, the copyright owner could seek orders requiring internet service providers, DNS providers, and—new and explicit in this bill—VPN providers to take “commercially reasonable steps” to stop their users in the United States from accessing those sites. The decision to label a website as a “foreign piracy site” can happen without the accused site even showing up in court to defend itself. 

ACPA Goes Further Than Other Site-Blocking Proposals 

In some ways, the ACPA is even worse than a site-blocking legislation introduced last year, the Foreign Anti-Digital Piracy Act (FADPA), which EFF also opposed. That bill at least excluded companies that provide only VPN services, as well as providers that offer DNS resolution exclusively through encrypted DNS protocols. The ACPA drops those protections. In fact, the bill explicitly includes VPNs among the service providers that can be ordered to block access to a website. 

The bill also broadens the definition of a “piracy site.” Last year’s site blocking bill covered sites with “no commercially significant purpose or use” other than infringement. ACPA changes that to sites with “only limited commercially significant purpose or use” beyond infringement. In other words, under ACPA, even a website with legitimate commerce going on could still be labeled a “foreign piracy site” and ultimately blocked for all Americans. 

Better Process Still Doesn’t Fix The Problem 

The ACPA includes some procedural protections, such as requiring service providers that could be subject to a blocking order to receive legal notice and an opportunity to respond. The bill also requires courts to consider the potential harm to other websites and internet users before ordering intermediaries to block websites. It further requires the copyright owner to post a bond, in an amount determined by the court, sufficient to cover the costs and damages incurred by any service provider found to have been wrongfully enjoined. The bill also provides a mechanism for operators or users of third-party online services affected by erroneous blocking to seek compensation after the fact in certain circumstances. Finally, a site operator can ask a court to rescind its designation as a “foreign piracy site.”

These safeguards are significant and positive changes, but they don’t solve the basic, and severe, due process problem. The initial decision to label a website a “foreign piracy site” can still be made without the site operator appearing to defend itself. The court can appoint a “special master,” which is an independent expert who helps the judge evaluate evidence, to review the copyright owner’s case—but that step is not required. In any case, a special master  is not a lawyer who actually represents the accused website, nor the users whose access to information and speech may be affected. 

We know what site-blocking looks like when it’s put into practice. Supporters of site-blocking like to point to its use in other countries. But what we’re seeing in other countries is serious collateral damage to lawful websites. In Italy, 510 benign, non-streaming websites, including a Catholic convent and a telehealth platform, were blocked by the country’s “Piracy Shield” program. In Spain, a site-blocking system blocked more than 550,000 domains during soccer broadcasts, including sites belonging to Greenpeace and Harvard University.

Congress Should Reject Site-Blocking Proposals

More than a decade ago, Congress abandoned SOPA and PIPA after internet users pushed back against site-blocking and other threats to the open internet. We shouldn't start building that infrastructure now.

ACPA adds some safeguards, but those don’t fundamentally change what Congress is being asked to create: a system for blocking Americans’ access to entire websites at the request of copyright owners. By explicitly bringing VPNs into that system, the bill also reaches into basic tools that people use to access the internet safely and privately. Adding somewhat better procedures to a bad idea doesn’t turn it into a good idea.

Victory! Appeals Court Rejects Expansive New Copyright Claim

 The U.S. Court of Appeals for the Ninth Circuit handed internet users and programmers a big win today, by rejecting an attempt to stretch a narrow provision of the Digital Millennium Copyright Act (DMCA) into a new source of copyright liability.  

The case involves Section 1202 of the DMCA, which prohibits intentionally removing copyright management information (CMI) like an author’s name or a copyright notice, from a copyrighted work. Open AI and Microsoft used code from Github as part of the training data for their LLMs, along with billions of other works. A group of anonymous Github contributors sued, alleging the new code coming out of these LLMs was similar to theirs—but with the CMI stripped out.  

The Ninth Circuit correctly agreed with what we said in our brief: removing copyright information from a copyrighted work is fundamentally different from creating a new work that didn't have CMI in the first place. Section 1202 of the Digital Millennium Copyright Act was intended to serve as a backstop for traditional copyrights in the digital age—not to create a new, more expansive right to inhibit otherwise non-infringing uses. 

As we also explained, accepting the Does’ theory would have created a brand-new source of liability for otherwise perfectly lawful activities, undermining creativity and innovation far beyond the specific context of AI development. Copyright holders would be able to file costly lawsuits against all kinds of legitimate users, such as artists making remixes based on older works, teachers adapting works for a classroom presentation, engineers reverse engineering code to understand it better, and search engines that help us all navigate the web. The risks would have fallen especially hard on independent software developers and other small creators. Large companies can afford to litigate these claims in federal court for years, if necessary. But an independent programmer facing massive statutory damages may simply have to settle, even when their underlying use is completely lawful. That’s why EFF fights to make sure courts don’t expand copyright beyond what Congress authorized.  

Copyright law still protects programmers when their work is unlawfully copied. They can still bring copyright infringement claims if someone uses a model to reproduce their code. Additionally, the plaintiffs’ contract claims against the AI companies are still in play. The specific holding here was narrow but important: that the absence of copyright information from a new work does not mean, by itself, that someone illegally removed it.  

That’s the correct result. New technologies will keep raising hard questions about copyright. Courts should answer those questions by applying the rights that Congress actually authorized, not by inventing new rights that could harm expression and lawful use for everyone.  

Additional Reading:  

Victory: Court, Using a New Test, Rules Embedding Links is Legal

Courts have for two decades found that linking and embedding someone else’s web content, be it a photo, music, or an article, doesn’t violate copyright law–the entity that controls the server that hosts a copyrighted work, not the user or website that merely directs others to it, is directly liable if the content turns out to be infringing.

News publisher Emmerich Newspapers sought to convince the Fifth Circuit Court of Appeals to chart a new and dangerous course, arguing that an aggregator website that published links to its copyrighted articles was in effect “displaying” them and can be directly liable for infringement. EFF, along with several other public interest organizations and trade associations, filed a brief urging the court to follow multiple other circuits and reject that theory.

Fortunately, the Fifth Circuit Court of Appeals did just that. While it rejected the server test–the rule courts have used to determine copyright liability rests with whoever serves up the content–the court came to the same practical conclusion by focusing on who is responsible for transmitting content. 

Applying that test, the court found that pointing or directing a user’s browser to request and receive the copyright owner’s own copy residing on its computers does not involve transmitting or communicating the content. “Although we take different routes to get there, both the server test and the test we announce end up in a similar place: a website cannot transmit a work that it does not have,” the court said. 

We told the court that accepting Emmerich's theory would make the common act of embedding links a legally fraught activity, one that many websites might be unwilling to risk, which would seriously damage the internet as a tool for creating and disseminating ideas and knowledge,

We applaud the court’s decision–even though it applied a different test, it correctly concluded that a user linking pictures, video, or articles isn’t in charge of transmitting that content to the world. The user doesn’t control what’s located on the other end of the link—that’s up to the person who controls the server.

Emmerich also claimed linking violates the Digital Millennium Copyright Act (DMCA), arguing its URLs were copyright management information (CMI) and when the aggregator displayed Emmerich’s articles under its own URL, it tampered with Emmerich’s CMI, which violates the DMCA. 

Under that logic, unsuspecting internet users could face ruinous legal risk for doing something as simple as using a link shortener, particularly given potential statutory penalties of up to $25,000 per violation.  

In our brief, we told the court that URLs don’t necessarily equate to a copyrighted work or provide sufficient information about the nature of the underlying content, making it highly unlikely that anyone would expect a URL to contain CMI. Quoting EFF’s brief, the court concluded that URLs are first and foremost a locational reference tool and while it may be possible for a URL to contain CMI, the bar to that conclusion is high.

Overall, this was a good and sensible decision that will protect ordinary online expression, communication, and access to knowledge. Hopefully this issue is laid to rest at last.

Court Rules Against Citizen Journalists in DMCA Takedown Case—EFF Will Appeal

A federal court in Massachusetts has ruled that copyright holders can issue online takedown notices based on a subjective belief of copyright infringement, even when that belief is unreasonable and self-serving. The case was brought by our client, Channel 781 News, after takedown notices temporarily shut down the citizen journalism group's YouTube channel. We think the court set the bar far too low for copyright takedowns, and we plan to appeal.

Channel 781 is a group of independent, volunteer journalists who report on local affairs in Waltham, Massachusetts. That includes posting short, newsworthy excerpts from recordings of city government meetings produced by Waltham Community Access Corporation (WCAC), the city's public access television station.

In September 2023, WCAC sent three copyright takedown notices to YouTube targeting fifteen of Channel 781's videos. YouTube removed the videos and, under its three-strikes policy, temporarily disabled Channel 781's entire account—just days before a local election.

Represented by EFF and Brown Rudnick LLP, Channel 781 sued WCAC under Section 512(f) of the Digital Millennium Copyright Act (DMCA), which provides a remedy when a copyright holder knowingly makes material misrepresentations in a takedown notice.

When Is a Copyright Holder Responsible for a Wrongful Takedown?

Fair use is the legal right to use copyrighted material without permission, when doing so serves purposes like criticism, commentary, or creating something new. Fair use is not copyright infringement, and courts have recognized that copyright holders must consider fair use before using the DMCA's powerful notice-and-takedown process.

In this case, Channel 781 argued that WCAC accused it of copyright infringement without making a good-faith assessment of whether its videos were fair use.

The evidence showed that WCAC's analysis was seriously deficient. The court noted that Chris Wangler, the WCAC employee who sent the notices, didn’t consider several facts relevant to fair use. For instance, Channel 781 used relatively small portions of WCAC's recordings, and the underlying recordings were factual public meetings, not a creative work. WCAC also gave little or no weight to whether Channel 781's use harmed any market for the recordings.

There’s also strong evidence that WCAC had motivations unrelated to copyright. WCAC objected to its footage being used to criticize local officials and advance political viewpoints. And WCAC sent the takedown notices during a local election, shortly after Channel 781 posted a campaign statement by Waltham's mayor that WCAC had mistakenly made available online.

Despite this evidence, the court concluded that WCAC had a subjective good-faith belief that Channel 781's videos were infringing. We disagree.

A Subjective Belief Should Not Be a Free Pass

Channel 781 argued that a copyright holder’s belief that material is infringing must be both genuinely held and objectively reasonable. WCAC argued that a subjective good-faith belief is good enough. Unfortunately, the court agreed with WCAC.

The court emphasized that Wangler had read up on fair use, watched a short YouTube video explaining the doctrine, and distinguished between videos he thought might qualify as fair use and those he believed did not. That was enough, the court concluded, to establish subjective good faith—even though Wangler’s analysis ignored important facts relevant to fair use. As the court put it, Section 512(f) does not require “a perfect or even reasonable fair use analysis.”

That is an alarmingly low bar for copyright holders seeking to remove someone else’s speech from the internet. A DMCA takedown can cause lawful speech to disappear almost immediately. As Channel 781 experienced, multiple notices can even result in an entire channel being disabled.

If a copyright holder can avoid liability despite a cursory, incomplete, and objectively unreasonable analysis that ignores important facts—even when there’s evidence that the copyright holder wanted to suppress critical speech—the obligation to consider fair use risks becoming little more than a box-checking exercise. That interpretation threatens to strip Section 512(f) of much of its force.

Even Under a Subjective Standard, WCAC Fell Short

Even accepting the court’s subjective standard, WCAC's cursory consideration of fair use should not have been enough. WCAC disregarded important fair use considerations, and the record included statements suggesting that it believed people generally needed permission to reuse its footage—an understanding at odds with fair use. There was also evidence that WCAC objected to Channel 781's political use of its footage, and had motivations for the takedowns unrelated to copyright.

Taken together, these facts raise serious questions about whether WCAC genuinely considered fair use, rather than using copyright as a rationale for removing material it did not like.

The Court Did Not Find That Channel 781's Videos Infringed

Importantly, the court's analysis recognized Channel 781’s strong fair use argument: the group used short excerpts from factual recordings of public government proceedings, selecting clips for their newsworthiness, and making them easier for the public and journalists to find, share, and discuss.

The opinion even states that WCAC's fair use analysis “may have been deficient.” But under the purely subjective standard it adopted, the court concluded that it could not reject WCAC's professed belief—even if the court itself “would have reached the opposite conclusion” on fair use.

We plan to appeal this decision to the First Circuit Court of Appeals. Copyright law should not allow a rightsholder to suppress critical reporting or political speech through the DMCA and escape accountability simply by claiming it believed the speech was infringing. Section 512(f) is supposed to provide protection against wrongful takedowns. We will keep fighting to ensure that safeguard actually protects people. 

EFF to Courts: Don’t Rewrite Copyright Over AI Hype

The history of technology is rife with copyright panics.  In the 1980s, major rightsholders ran to Congress and the courts, claiming that videotape recorders (VTR) were “to the American film producer and the American public as the Boston strangler is to the woman home alone.” Then, the Supreme Court declined to embrace the hype, noting that the VTR was capable of all kinds of non-infringing uses, like time-shifting and cautioning courts to avoid rewriting copyright law in response to new technologies. We believe that courts now should be similarly wary about the hype surrounding AI.

Hollywood’s hyperbole has echoed that of composer John Phillip Sousa, who claimed in 1906 that the player piano and the gramophone would destroy music composition; portrait artists who feared the camera would replace the paintbrush. None of these things happened. Cameras, for example, sparked a resurgence of portraiture and, by making it possible for more people to create images, led to unexpected developments—like the rise of photojournalism.

New markets, new ideas, and new creators are actually what copyright is supposed to promote, not restrict. Using copyright to lock in existing gatekeepers and massive rightsholders’ profits helps neither the public nor individual artists.

Generative AI has sparked the latest wave of anxiety and with it a massive wave of litigation. In multiple cases around the U.S. and the world, rightsholders are asking courts to do precisely what the Supreme Court warned against: dramatically expand copyright protections based in substantial part on hyperbole and speculation. They should decline to do so.

Copyright owners claim that unless courts abandon 300-year-old copyright principles—and give rightsholders the power to control non-infringing works created by others—an imagined flood of AI-generated works will devastate creative markets. Under this “market dilution” theory, building generative AI tools cannot be fair use because those tools might be encourage the proliferation of competing works.

As EFF has explained to the courts in multiple amicus briefs in Concord Music Group, Inc. v. Anthropic PBC and In re Mosaic LLM Litigation, that’s not how copyright works. In fact, accepting this theory would undermine copyright’s constitutional purpose: promoting the creation of expressive works for the public’s benefit. Because copyright law is designed to encourage others to build freely on existing works, it punishes infringement, not competition. The “market dilution” theory would eviscerate not only the fair use doctrine, but also other limits on copyright that work specifically to prevent rightsholders from unfairly suppressing competition by claiming broad ownership over tropes, genres, styles, and so on. In other words, publishers would wield unchecked veto power over any expression that might conceivably compete with a work they own.

The result? Art doesn’t get created, ideas are never expressed, and we’re all worse off. Copyright shouldn’t be a tool to silence future creative competitors—whether or not they use AI in their work.

And the plaintiffs in these cases get at least two other things wrong. First, research shows that large generative AI models are unlikely to produce infringing works because the more data on which a model is trained, the less any individual training example matters to any particular output.

Second, AI tools aren’t necessarily displacing human creativity. To take a just a few examples:

  • Boston-based artist Nettrice Gaskins uses AI to create Afro-futurist art, including a portrait of Octavia Butler displayed at the San Francisco Airport
  • Indian artists Prateek Arora and Varun Gupta use generative AI to reimagine Western science fiction.
  • Philadelphia-based artist Alex Smith uses generative AI to reimagine Afrofuturism with queer, plus-sized Black superheroes.
  • Ana Miljački, a professor of architecture at MIT, used generative AI to create a “non-liner documentary” film on Yugoslav World War II memorials and the values they embodied.
  • A research-creation project used AI generated visual art to both amplify the voices of activists in the Iran Woman Life Freedom Movement and evaluate AI’s role in sociopolitical advocacy through art.
  • AI company Bronze works with musicians like Disclosure and Jai Paul to create songs that never sound the same when played back twice, challenging audience conceptions of what music could be.

It is not the place of courts to say these people are not artists or that AI cannot augment human creativity in a positive way.

Given this range of experimentation, courts should be reluctant to decide in advance what tools do and do not foster “human creativity.” Like the VTR, large language models are general purpose tools, used by humans to do a broad variety of things far beyond generating lyrics. The effects of this particular technological innovation will doubtless be far-reaching, disruptive, and potentially harmful for some—but distorting copyright law is not the way to address those harms.

The CHATBOT Act Forces One Parenting Model On Every Family

Artificial intelligence is rapidly changing education, and the way people search for information. Parents, teenagers, teachers, and schools are struggling with tough questions about when AI should, and should not, be used. It makes sense for Congress to hold hearings and examine how AI should be used by minors. But the recently introduced CHATBOT Act answers those questions with a one-size-fits-all mandate governing how teenagers access AI through federally prescribed parental monitoring systems. 

The Bill Requires AI Companies To Build Family Monitoring Systems 

Parents are approaching AI in different ways. Some closely supervise how their children use chatbots, while others might set more general rules about technology. Many families are still figuring out what role AI should play in schoolwork and everyday life. 

The CHATBOT Act would take that decision away from families and AI providers. Instead of letting families and AI providers decide what parental controls should look like, Congress would require every covered AI chatbot to build the same federally prescribed “family account” system. 

As part of the required parental-consent process for teens, AI companies must offer parents a "family account" that provides access to a "full record of the conversations and activity" of teen users and tools to "monitor, analyze, and understand, at scale" those conversations. They must also send alerts if a teen attempts to bypass or disable parental controls. 

This isn’t simply an optional parental-control feature. The bill requires every covered AI provider to build this monitoring infrastructure, and present it as part of the parental consent process. Congress is prescribing a single, highly invasive model of how families should supervise teenagers’ use of AI. 

The CHATBOT Act Creates New Privacy Risks For Families 

Parents and families have different ideas about how much independence teenagers should have. Understandably, they also have very different expectations for 8-year olds, 13-year-olds, and 17-year-olds. The CHATBOT Act effectively requires AI providers to build the same monitoring architecture for users of very different ages. 

And this mandated data collection will create new privacy and security risks. Once Congress requires AI companies to create a permanent, centralized record of teen AI conversations for parental review, that will be a valuable vault of extremely personal information. That raises serious questions about what would happen in cases where someone else gains access to it through account compromise, family disputes, or other security failures. 

The vast archives of conversations created by the government-mandated family accounts won't be interesting only to parents. They will become valuable targets for hackers, identity thieves, civil litigants, and anyone else seeking access to the deeply personal information of others. The CHATBOT Act requires the records to exist, but addresses none of those risks. 

Families are still figuring out what role AI should play in schoolwork and everyday life. Congress shouldn’t freeze one answer into federal law by requiring every AI company to build the same prescribed monitoring system. 

The CHATBOT Act Applies A Children’s Law To Teenagers 

The CHATBOT Act takes the basic structure of COPPA, a nearly 30-year-old law that applies to children aged 12 and under, and applies the same “verifiable parental consent” to older teenagers. 

That’s a dramatic expansion of the law. Congress enacted COPPA to prevent kids from handing over detailed personal information to online services without making sure parents approved. For nearly three decades, Congress has required parental consent before websites collect personal information from any user under 13. COPPA is not simple to comply with, which is why so many internet companies, large and small, simply bar kids under 13 from having accounts. That includes major social media sites and AI. Facebook, Instagram, TikTok, X, YouTube, Snapchat, Discord, Spotify, and blogging platforms like WordPress all keep out users under 13. Children under 13 are also not allowed to use Microsoft Co-Pilot, Google Gemini, or ChatGPT. Anthropic does not allow users under 18 to use its AI model, Claude. In cases where younger kids maintain social media accounts despite the rules, studies show the vast majority of them are creating those accounts with parental consent.  

In short, COPPA’s protections against collecting personal information from minors without parental consent already apply to the AI services CHATBOT Act seeks to regulate. Worse, the CHATBOT Act takes COPPA’s privacy protections and inverts them—it will result in AI services likely collecting more information about young users. 

But the CHATBOT Act extends that model to high school students using AI assistants that are rapidly becoming tools for learning, research, writing, coding, and creative work. It then mandates specific, invasive surveillance tools that go well beyond anything COPPA requires. 

The bill requires providers to offer these “family accounts,” with these specific features, as a default for teenagers. By doing so, CHATBOT effectively treats a high school senior the same way it treats an elementary school student. 

Supporters may argue that parents of teens don’t have to create a family account. But every family with a teenager will still have to go through the bill’s parental-consent process before a teenager can use a covered AI system. Providers will need practical ways to verify that an adult is, in fact, the teenager’s parent. And parents of kids under 13 have no option to consent to their kids’ use of an AI system—the bill’s only option is to create a family account.

Congress should not extend the COPPA parental-permission model to millions of older teenagers, and it would be harmful to do so. The government does not require COPPA-style parental permission before a 17-year-old checks out a library book, uses Wikipedia, types search terms into Google, or reads a newspaper online. It shouldn’t require parental permission simply because the same question gets asked of an AI assistant. 

The CHATBOT Act Will Pressure AI Companies To Check Users’ Ages 

The bill says it doesn’t require age verification. But like many recent “kids online safety” bills, it imposes obligations that depend on a company knowing whether a user is under 18. 

Specifically, the bill requires AI systems to either disable access to young kids, get parental consent, or the creation of a family account if a service has reason to believe a user is a minor. The standard means that services don’t need to have actual knowledge of a user’s age to be later held liable for improperly letting them use their AI tools. That creates a practical problem. Given the potential liability of getting something wrong, AI companies will likely require stricter forms of age verification to figure out who is under 13, a teenager, and who is a parent. Some providers might ask for government-issued identification.  Other companies may rely on age estimation systems that use facial scans or other signals to guess a user’s age. Neither of these approaches is good for users’ privacy or security. One collects more information than is necessary, and the other inevitably makes mistakes. 

Congress shouldn’t force companies into that choice, or families into this position. In the name of protecting children, the CHATBOT Act will result in online services collecting even more information from kids and families, creating privacy and security risks. Parents who want family accounts like those described in the bill should be free to choose AI services that offer them. But Congress shouldn’t pressure every provider to collect more information about everyone’s age simply to comply with the law. 

A Better Way Forward

Congress doesn't have to choose between doing nothing and creating a sweeping new federal parental-monitoring mandate. Existing law allows regulators to police deceptive AI products, protect children's privacy under COPPA, and hold companies accountable when they market unsafe or misleading products to families. 

Lawmakers have urged the FTC to crack down on AI-enabled toys that make unsubstantiated educational claims or illegally collect children's data. Those are regulatory actions that can be taken right now. 

Finally, the FTC is currently investigating how AI companies test their products, protect children and teens, comply with COPPA, and enforce age restrictions. The results of that inquiry could be useful guidance to Congress, and to the public debate around these issues. 

Cracking down on bad actors, while learning more about how families are already making decisions about AI use, is a much better path forward than building one, federally-prescribed model of parenting or product design.

Don’t Repeat NY’s 3D Printing Blunder

This year the state of New York had the dubious honor of being the first to pass a controversial provision to mandate all 3D printers come with surveillance and censorship. That means not only is there a ticking clock to protect every artist, researcher, engineer, and hobbyist in the state, but there is a real risk of other states thoughtlessly following suit—prior to the New York rules even taking effect.

We, along with many other experts, already warned about this bill buried in the state’s crowded budget process. Hundreds of our supporters and 3D printing enthusiasts in New York reached out to their representatives hoping to kill this farcical bill. While there were some welcome amendments in response to the outcry, Albany passed it anyway.

It might be well-intentioned, but bills like these sell a fantasy that can only have an untold negative impact on the privacy, free expression, and consumer rights of anyone using these general purpose devices. Behind the banner of reducing gun violence, which is nearly always committed with commercial firearms, New York lawmakers have passed draconian legislation that will let manufacturers lock in users and collect their data.

Now that the bill has passed and been signed by Governor Hochul, let’s look at two important ways the final legislation changed since we last wrote about it, and why states like California shouldn’t make the same mistake.

Reduced Risk for Lawful File Sharing 

The New York bill includes language that criminalizes access to firearm print files, a proposal correctly dropped by states like Colorado due to First Amendment concerns. While this made it through to the passed legislation, a few wins were still gained.

Originally the legislation threatened felony charges for the storing and sharing of files, potentially impacting researchers, artists, and journalists with no intention of printing a firearm component. These charges were downgraded to a Class A misdemeanor.

Two provisions criminalized file sharing. The first of the two provisions criminalizing this file sharing, which pertains to the sale or distribution of files in the state, gained an important exception for when a sender has a reasonable belief that the recipient won’t illegally print these components. However the second provision, pertaining to criminalizing file possession, complicates this. Under 2.12 of the subpart, people who possess the file with intent to share the files do not clearly get this same reasonable belief exception.

In other words, if you share one of these files the actual sharing is covered by the exception, but the law makes it ambiguous whether possessing those same files is covered when you intend to share them.

While this exception could have created some breathing room for researchers and journalists operating in good faith, this slapdash bill language leaves plenty of ambiguity and potential speech-chilling effects. However, these changes do offer a modicum of harm reduction in this unconstitutional law.

Saving Face by Preserving Online Sale

Originally the bill had a strange requirement for all 3D printers and Computer Numerical Control, or CNC, machines to be sold and delivered face-to-face, with no exception. That would have meant a major barrier to access, particularly for people in agricultural and rural areas of the state who uniquely benefit from in-home fabrication and repair. It also would have meant a major inconvenience for businesses using these devices. For everyone though, it meant fewer retailers to choose from and facing more stigma for using these devices. 

Fortunately this was dropped from the bill entirely. 

Next Step: We Find Out What Was Actually Passed

In addition to being buried in the complicated legislative process of the NY budget and avoiding proper scrutiny, this bill also kicked the can down the road in determining what exactly is being mandated. In many respects, legislators passed a vibe. We’ll see how the actual law be developed over the next year by a working group with no mandated transparency to the public. Further, they have no obligation to ensure consumer safeguards in developing this state-mandated censorware.

We are still concerned by the possibility of a biased working group acting in the interest of manufacturers or facing pressure to accept consumer harms in the standards they produce. Our remaining hope is this working group convened by the Department of State and the state university system is composed of actual experts who are aware of how unfeasible and harmful this mandate is, and prevent it from being realized.

The Fight Continues

New York is the first to go down this path of state-mandated censorship and surveillance software on 3D printers, but it’s far from the only one to entertain it. It is now more urgent that we fiercely oppose this trend in other states, like California,  as they attempt to join the bandwagon—before even seeing the real-world impacts. 

Take action

Don’t Let California Repeat NY’s Mistake

We cannot allow this to be the foundation for future restrictions on speech and design, or serve as a playbook for the state and corporations to wrest control over our tools.

Sony Nerfs Videogame Ownership

Legal intern Suzanne Castillo co-authored of this post.

Playstation’s decision to kill physical game discs is the latest attack on our diminishing rights to access and engage with culture digitally. Rent-seeking corporations and negligent lawmakers share the blame — and they can do better. 

We’ve seen the same playbook used in the move to digital distribution of  film, TV, and music: draw in customers with the convenience of a digital download, then limit physical access and move the goalpost on what it actually means to “own” a piece of media. The end goal is to turn the customer into a renter, stuck making regular subscription payments for access. Gamers are right to sound the alarm, and we must take this moment to fight for digital ownership before it’s too late.

Disk Space Invaders

Depriving gamers of physical discs leads to another obvious and immediate cost: data.  Unlike other digital media like film and TV, video games require a ton of storage. Access to high speed internet is still abysmal in the US, making the high-speeds needed for digital game downloads a luxury some of us may take for granted. For many, a modern game can take days and exceed their data caps. 

This made physical discs, particularly for the biggest AAA titles, a logical choice that also largely spared gamers from losing traditional ownership rights. With physical disks, the cost of storing the game was included in the purchase.

Own or Be Pwned

Limiting customers to digital copies also pushes gamers further into rent-only copyright culture.

Physical media comes with a "right of first sale," which means you can lawfully share, resell, alter, or destroy your own copy of a copyrighted work. This right has also helped protect the emergence of alternative community servers, and emulator addition of online play to games from the dial up era.

But courts have held that digital media doesn't carry the same right, meaning no such protection is afforded to digital purchases. Your ability to freely share games with friends or pass them on to family members becomes totally subject to the whims of the distributor. 

So, for example, a digital-only approach effectively guts the second-hand market for games. Saving some money with a used game and recouping the costs by reselling are no longer an option. Even with steep discounts and holiday sales, this raises the minimum cost of engaging with the medium at all.

The inevitable conclusion of the move to digital-only purchases is to lock gamers into  subscription models, making their access totally dependent on the distributor — or, several distributors, as we’ve seen with major TV and movie streamers. A handful of companies actually own the games, and your only option is to regularly pay for fractured libraries of games you may never play and will never truly own.

Achievement Locked

Since digital games are easy to copy, distributors and publishers argue that they are in an arms race against piracy. The irony is that law-abiding customers consistently suffer collateral damage. 

Most digital distributors lock down the content they offer with restrictive user agreements and digital rights management (DRM) software. DRM software, in particular, imposes onerous controls on the game — like forcing internet connection for single player games or modifications that harm performance — and can even introduce serious privacy and security concerns. Any gamer or researcher in the US who wants to reduce this burden by removing or modifying that DRM risks a lawsuit, thanks to Section 1201 of the Digital Millennium Copyright Act (DMCA). This federal law makes it illegal to alter DRM software, and is a beloved tool for companies trying to restrict how we can lawfully use our purchases — whether it’s a copy of the newest tractor simulator or a literal tractor. 

And since much of this DRM is tied to user accounts, ownership of a game is also revocable and modifiable for any number of reasons outside of your control. Error in your subscription payment? Your account got hacked? Licensing deal falls through with a major publisher? Developers want to kill the game in an update? All of this can limit or change your ability to access the game long after your so-called “purchase.”

Level-up Ownership

Policymakers can and should work to restore our ownership rights for the digital age. 

That starts with legal protections ensuring that the same rights that apply to physical media apply to digital media. Next up? Reform Section 1201 of the DMCA to clarify that it does not forbid fair uses.  

At the state level, we need meaningful consumer protections. Some promising models include California’s AB 1921, which would clarify what customers are actually paying for on digital storefronts and ensure some protections for maintaining discontinued games. The gaming industry has done its best to kill the bill, including claiming that private community servers are illegal. 

If you bought it, you should own it, and EFF will continue working to mitigate some of the worst harms of the DMCA 1201, defending modders, and fighting deceptive licensing that makes culture less free.

We Can Still Stop California’s 3D Printer Surveillance Scheme

Ignoring EFF’s warnings about the dangers and impossibility of implementing a new mandate for 3D print surveillance software, the California State Assembly has signed off on legislation to do just that. In the process, legislators amended the bill to make it even more confusing, while failing to address the risks to privacy, speech, and consumer rights. We must renew our call on legislators to drop this bill as it heads to the state senate, and protect the tools of creators in the state.

Take action

Tell CA Senators to stand with creators

What’s changed about the bill?

Since we first wrote about AB  2047, a bill targeting 3D printers for the rare, impractical, and already outlawed practice of manufacturing firearms without a license, it has picked up several amendments. Some are welcome changes, but most have only highlighted the technocratic absurdity of the proposed scheme. Our core concerns—that this mandate censors lawful speech, builds out corporate surveillance, and criminalizes open source experimentation—have not been remedied. 

Removes criminalization of resale

Starting with one silver lining, the current bill includes a carveout for the private resale of devices. The original bill would have made it a criminal offense for an individual to resell 3D printers purchased before this mandated censorship and surveillance software. This is a clear win for the 3D-printing community, but it is unfortunately not enough.

Ineffective carveouts for open source

One of the most dangerous aspects of the bill is that it criminalizes individual users for common practices, like creating and using alternative open source programs with their 3D printer. New amendments provide a carveout for the use of an open source tool, but only if it includes compliant censorship software. The bill burdens open source developers with ambiguous and unrealistic standards for print blocking, and continues to create a chilling effect for open source users.

Removes any actual requirement to work

To reiterate—there is no world where the mandated technology actually works as intended. It will both block lawful use of 3D printers, and allow firearms to be printed by anyone determined to do so. There is no amendment that can change this reality.

Instead, the current bill simply drops the pretense that this mandate is expected to work. The performance standard of algorithms changed from “effectively prevent[ing] a technically skilled user from evading [the algorithm]” to “substantially reduce the likelihood of foreseeable circumvention attempts…” The bill will still require all prints to be surveilled, but instead of testing efficacy against a skilled user, it just plays whack-a-mole with the (literally) infinite number of circumventions that any user can employ. 

Further, the bill now leaves us with an unclear process that relies on non-governmental third parties to define standards, and now relies on manufacturers and resellers to self-police.

Hollywood gets a cut

The bill includes yet another carve out for commercial users. This time for the entertainment industry, which makes extensive use of 3D printers for props and costumes. 

That’s fine for big studios, but it leaves out indie filmmakers, cosplayers, and many other small creators. 

This is simply a defensive edit to limit corporate opposition. There isn’t a clear division in 3D-printing between consumer and commercial tools. These are general purpose tools which might be picked up by a prop department of a big studio, or an artist getting ready for Comic Con. Indeed consumer level products are not only used by amateur artists and engineers developing their skills. Commercial 3D printers, like their traditional 2D equivalents, are frequently used in workplaces, as well as by professionals honing their skills or just trying to get some work done at home. 

Commercial carveouts hands printer manufacturers the ability to sell a more expensive tier of printers, locking-in and up-charging their commercial customers. Some of those customers will choose to buy general retail versions, but that carries its own price: increased risk of IP theft as all printed files are surveilled the same way they are for hobbyists. That means a real risk of businesses leaking any prototypes or new designs to not only the printer manufacturer, but potentially snooping governments and/or the general public through data breaches.

Demand  your senator oppose AB 2047

This updated version of AB 2047 downgrades performance standards and removes oversight while still threatening privacy and choice for users of 3D printers. A printer surveillance system won’t work for its intended purpose, and will only harm law abiding users. 

Act now to demand your senators to vote no on this ineffective and invasive bill.

Take action

Tell CA Senators to stand with creators

The NO FAKES Act Could Silence Satire, Commentary, And News

The NO FAKES Act is supposed to target harmful AI-generated impersonations. But in reality, it will make it easier to suppress commentary, satire, and other lawful speech. That's why EFF has signed a letter urging the Senate Judiciary Committee not to advance the bill in its current form.

Take action

Tell Congress to Say No to NO FAKES

In the letter, EFF joins a coalition of civil society groups in pointing out that the bill would import many of the worst features of the DMCA notice-and-takedown system into an even broader range of online expression. Faced with a “heckler’s veto” over legal speech, platforms will have incentives to remove content first and ask questions later. 

The bill offers no protection for a platform’s judgment about an often difficult question—whether a particular piece of content is satire, parody, commentary, or news. Any platform that guesses wrong faces penalties of up to $750,000 per work. 

NO FAKES could also undermine the rights of the people it is supposed to protect. The new federal “likeness” right could be licensed or transferred to others, so individuals will lose control over the use of their own face and voice. That’s not theoretical—workers in the entertainment industry are routinely asked to sign broad contracts about the future use of their likenesses.

As the letter notes: 

A background actor who signs a release on set or an ordinary person who clicks through a platform's terms of service could end up with the right to their own face and voice in someone else's hands, for years, with federal enforcement behind it. 

EFF and the other signatories urge Congress to examine existing legal remedies and pursue narrowly tailored solutions to genuine harms. The last thing we need is a sweeping new intellectual property right that threatens free expression. 

In addition to EFF, the letter is signed by the Center for Democracy & Technology, the American Civil Liberties Union, Fight for the Future, Foundation for Individual Rights and Expression, the Organization for Transformative Works, Public Knowledge, the R Street Institute, The Future of Free Speech, and the Woodhull Freedom Foundation. Read the full letter here. 

Take action

Tell Congress to Say No to NO FAKES

Congress Just Rushed Through a Disastrous Copyright Office Overhaul

In a voice vote earlier this week, the House of Representatives passed H.R. 6028, the “Legislative Branch Agencies Clarification Act.” The legislation is presented as a technical reorganization of some government agencies, but it’s much more than that. 

H.R. 6028 would fundamentally change the U.S. Copyright Office, and not in a good way. The bill removes the Library of Congress’ current supervisory role over the Copyright Office, transfers several powers directly to the Register of Copyrights, and makes the Register a presidential appointee, confirmed by the Senate. 

These changes would make an office that’s already hugely influential in copyright and tech policy much more political. EFF first explained why that’s a terrible idea when it came up nearly a decade ago. This bill, like the older one, weakens the few public-interest checks and balances that do exist.  We hope the Senate promptly rejects this bill. 

The Copyright Office Doesn’t Need More Politics—Or More Power

The Copyright Office's main responsibilities are administrative and advisory. It registers copyrights, maintains records, grows the Library of Congress’s collections, and provides expertise to Congress on copyright law. But over the past two decades, the Office has also become increasingly influential in copyright policy debates that affect free expression, libraries, educators, competition—and everyday internet users. Unfortunately, it has not been a neutral advocate. The office’s recent report on the role of AI severely bungled the issue of fair use, prioritizing private licensing market “solutions” over user rights. 

Going further back, the Copyright Office supported one of the most infamous anti-internet proposals of all time—the Stop Online Piracy Act (SOPA), a disastrous internet censorship proposal that sparked one of the largest online protests in history. The Office has repeatedly advanced positions that favored large entertainment-industry interests over the public interest.

The Office also plays a major role in the Digital Millennium Copyright Act (DMCA) Section 1201 rulemaking process, which determines when the public may lawfully bypass digital locks for activities such as security research, repair, preservation, or accessibility. EFF has used this process repeatedly to mitigate some of the worst harms of the DMCA. H.R. 6028 would move rulemaking authority over 1201 from the Librarian of Congress to the Register of Copyrights, further consolidating power within the Copyright Office itself.

The bill also makes the Register of Copyrights a presidential appointee confirmed by the Senate. Each administration will be pressured to pick nominees aligned with their own policy preferences, and the powerful copyright owning industries will invest even more heavily in lobbying to get their way, and influence the selection. This position should be focused on administrative ability and actual expertise, not lobbying and politics. 

The Copyright Office Should Stay Connected To The Library of Congress

H.R. 6028 would do more than change who appoints the Register of Copyrights. It would sever the Copyright Office from Library of Congress supervision and transfer many Librarian powers directly to the Register. 

The supervisory relationship exists for good reason, as the nation’s libraries have pointed out for years. The Library, while far from perfect, at least has the mission of preserving and providing access to knowledge. That should be an important public-interest counterweight in copyright debates. Congress has not explained how weakening the ties between the Library and the Copyright Office would serve the public better, or even seriously inquired about it. 

This Bill Was Rushed Through

Back in March, EFF joined Public Knowledge, the Center for Democracy and Technology, library organizations and tech groups, urging Congress not to fast-track this legislation. We told them changes to the Copyright Office will have major consequences for the “speech rights, educational opportunities, and creative freedoms of all Americans.” 

Yet Congress moved forward without any hearings on the bill, and without meaningful examination. H.R. 6028 creates a years-long separation of the Copyright Office from the Library of Congress, transfers significant legal authority, and restructures the appointment process for the nation’s top copyright official. Changes like that deserve hearings, debate, and public scrutiny. H.R. 6028 got none of that. 

The Senate Should Stop This Bill

Copyright law exists to serve the public and “promote the progress” of science and learning. The institutions that administer copyright law should do the same. 

H.R. 6028 would move the Copyright Office further away from that goal. Congress should be strengthening public-interest oversight of copyright policymaking, not looking for ways to concentrate more authority in a single presidentially appointed official. 

The Senate should reject H.R. 6028. The Copyright Office should serve the public—not presidential administrations, and not industry lobbyists. 

❌