Paul Keller
@paul.eurosky.social
Director of Policy at Open Future, President of the COMMUNIA association for the Public Domain. Advocate for more open, just & inclusive digital policies in 🇪🇺 & beyond. Collector of collateral knowledge, © nerd, cyclist.
More control for rightsholders or easier access for AI developers? Only shifting the obligation from training to deployment can deliver both. Analysis based on 432 responses to the EU Commission’s call for evidence regarding a new legislative initiative on copyright. openfuture.eu/blog/squarin...
One last thing before I log off for the summer break: my analysis of all 432 responses to the EU consultation on AI and copyright. Nearly every submission asks the Commission to deliver one half of its dual objective at the expense of the other. A small group shows how the circle could be squared.
New from 404 Media: a company is selling tons of old books specifically to AI companies and says they have a strong NDA to never reveal which AI company bought all of the books. They don't want headlines of 'we scanned and destroyed millions of books for AI' www.404media.co/ai-companies...
AI Companies Are Buying Tons of Old Books Because They're Free of AI Slop
ISBNdb, a company that sources printed books for AI companies to turn into training data, tells clients “the optics problem is real.”
404media.co
Who pays for the commons? Generative AI was built on the open web, but little value has flowed back. @paul.eurosky.social’s latest at @botpopuli.bsky.social argues copyright can't fix this on its own — and makes the case for a levy on commercial AI deployment. Read it: botpopuli.net/who-pays-for...
Who Pays for the Commons?
Three and a half years after the emergence of generative AI as a new technology paradigm, there is broad agreement that AI companies have extracted enorm...
botpopuli.net
Very happy to see our Feasibility Study for a European Books Data Commons published. @openfuture.bsky.social has worked with @europeana.bsky.social, @kb.nl & others across the European library community on one question: can Europe's libraries build a single access point for public domain books?
The EU wants digital sovereignty built on open source AI, but its own copyright rules stand in the way. A new study by @communia.bsky.social with Maria Drabczyk (Centrum Cyfrowe) & @alek.eurosky.social maps the legal friction and how to fix it. openfuture.eu/publication/...
Copyright challenges in open-source AI development in the European Union – Open Future
The EU wants to build digital sovereignty through open source AI, yet its own copyright rules stand in the way. A study of eight European initiatives maps the resulting legal friction and proposes sol...
openfuture.eu
A copyright exception you can't use isn't really an exception. That's the core of our input to the EU copyright review: close the gap between what the law lets libraries and archives do and what they can do in practice. www.internetarchive.eu/2026/07/01/r... #digitalpreservation #openaccess
Rights on Paper Are Not Enough: Our Input to the EU Copyright Review - Internet Archive Europe
A copyright exception you cannot use is not really an exception. That idea runs through the submission Internet Archive Europe (IAE) submitted to the European Commission this month. On 25 June, we res...
internetarchive.eu
The Dutch GPT-NL project shows that refusing data permitted under EU law — in the name of "respecting copyright" — holds the model back. Sustaining Europe's AI ecosystem means rebuilding copyright as a right to be paid. Read more in @paul.eurosky.social's latest: openfuture.eu/blog/gpt-nl-...
GPT-NL respects copyright — cui bono? – Open Future
GPT-NL refuses training data that copyright law permits. The result is a model that cannot compete, highlighting why sustaining the information ecosystem requires the right to be paid rather than a ri...
openfuture.eu
The EU has pledged €200 billion for AI. How much of that reflects actual commitments, and how much is political ambition? Our latest report, "From Frameworks to Factories", provides a baseline for analysis: openfuture.eu/publication/...
Unconditional openness doesn't protect the commons when AI companies extract value from public collections at scale. It enables privatization. Analysis by @paul.eurosky.social responds to critiques of our proposed conditional access framework for cultural heritage data: openfuture.eu/blog/beyond-...
Beyond Control and Compensation – Open Future
Conditional access to cultural heritage data isn't about control. It is about reciprocity in a context where public collections are being extracted on an industrial scale.
openfuture.eu
so this weekend’s news reminds me of the fact that (a) they were selling #usaf #f15e branded chocolate bars in iran on 2007 and that (b) i was somewhat competent at predicting the future back then flic.kr/p/49cKrK
Anata F-15 chockolate bars (defiance!)
chocolate bar display box on the counter of a juice stall on the main street in Shiraz. kind of ironic that they put pictures of the same planes on chockolate bars that the US and the Israelies would ...
flic.kr
🚀 Registrations are now open for the 2026 Institute for Information Law (IViR) Summer Course on International Copyright Law & Policy! Join us for an intensive 5-day program on international copyright law, featuring an incredible lineup of experts 🔗 Details & registration: www.ivir.nl/courses/icl/
Summer Course on International Copyright Law & Policy 2026 - IVIR
The Institute for Information Law organizes an annual intensive post-graduate course on international copyright law and policy. The face of copyright law and policy changes continuously, mostly due to...
ivir.nl
well that was very uneventful @paulk.bsky.social is now @paul.eurosky.social. #eurosky (I guess uneventful is good in this context, so 👏 @eurosky.social)
Publishers blocking @archive.org crawlers over AI concerns. Reporting from @niemanlab.org shows how scarcity-based responses to AI are eroding open web infrastructure in real time: www.niemanlab.org/2026/01/news-p… I wrote about why this path leads nowhere good: www.techpolicy.press/abundance-vs-s…
News publishers limit Internet Archive access due to AI scraping concerns
Outlets like The Guardian and The New York Times are scrutinizing digital archives as potential backdoors for AI crawlers.
niemanlab.org
Don't forget to check the CommonsDB project, a public domain and openly licenced works registery www.commonsdb.org
CommonsDB
We are building a public registry for Public Domain and openly licensed works
commonsdb.org
Interested in the European Commission's consultation on TDM rights protocols closing 23 January? Our new policy brief by @paulk.bsky.social examines how AI opt-outs evolved in 2025, proposes four principles for a shared vocabulary layer, and provides timely context. openfuture.eu/publication/...
in the midst of all the chaos I took the train to Warszawa and arrived right on time: blog.voyantes.net/kind-train-g...
meanwhile...
blog.voyantes.net
The proposed Digital Omnibus would allow public bodies to charge tech giants different fees for reusing public data. @paulk.bsky.social examines where this works—and where it adds unnecessary complexity—addressing the Paradox of Open and its implications for open data. openfuture.eu/blog/opennes...
Openness, Power, and the Paradox of Open Data in the Digital Omnibus – Open Future
The Digital Omnibus goes beyond technical fixes—the EU is finally acknowledging the Paradox of Open. Differentiated charging makes sense. Weakening standard licensing risks fragmentation.
openfuture.eu
when all of NL is in a #winter meltdown, but you have to get to Warsaw and the train gods like you. Bravo @deutschebahn.com!
The new Agentic AI Foundation (OpenAI, Google, Microsoft...) will set rules for how AI agents access data. Wikimedia's voice is missing. If this is about the public interest, the digital commons must have a seat at the table. More in the latest from @alek.bsky.social: openfuture.eu/blog/why-wik...
This new impulse paper proposes a framework to help cultural heritage institutions balance their commitment to open access with the increasing AI-driven demand on their collections and infrastructure. ↘️
Open Future@openfuture.bsky.social · 8mo ago🏛️How should cultural heritage institutions share data in the age of AI? Our new paper with @europeana.bsky.social explores when heritage data should be available for AI training. Part of the Alignment Assembly on Culture for AI. Read more: openfuture.eu/publication/...
@paulk.bsky.social Keller @openfuture.bsky.social read the GEMA v. OpenAI judgment. In his view, the judgment reaffirms the TDM exceptions enable AI training—provided the training doesn’t lead to memorization. openfuture.eu/blog/g... #AI
GEMA v OpenAI: Memory is Fragile. Garbage Lasts Forever. – Open Future
GEMA's lawsuit against OpenAI has reignited the debate about AI training and copyright. This post examines the relationship between AI training and the TDM exceptions.
openfuture.eu
🇪🇺 The EU has the legal tools to build public AI—but uncertainty is holding it back. Article 3 of the CDSM Directive lets research organizations & libraries train models on accessible works. Read @paulk.bsky.social’s latest analysis for more: openfuture.eu/blog/leverag...
Leveraging Article 3 of the CDSM Directive to build public AI models – Open Future
Article 3 of the CDSM Directive allows EU research organizations and libraries to train AI models on copyrighted works. This analysis explores how this legal framework enables public AI development.
openfuture.eu
First EU ruling on AI training: TDM exceptions apply, but models can't memorize copyrighted works. Court draws line between learning and leakage, leaving key questions on opt-outs and memorization thresholds. Read @paulk.bsky.social latest: openfuture.eu/blog/gema-v-...
GEMA v OpenAI: Memory is Fragile. Garbage Lasts Forever. – Open Future
GEMA's lawsuit against OpenAI has reignited the debate about AI training and copyright. This post examines the relationship between AI training and the TDM exceptions.
openfuture.eu
I have been a paid subscriber to @404media.co since pretty much day 1. Easily the best $100 you can spend on the internet. Yesterday I accidentally heard their podcast with ads—I’d probably pay $100 just to avoid those cringe-worthy things alone.
📝 Voss's draft report on copyright & AI shows irony: questions his own framework only to circle back to it. His focus on remuneration and transparency offers pathways to a more sustainable information ecosystem. Read @paulk.bsky.social’s analysis: openfuture.eu/blog/licensi...
Licensing, Levies, and the Limits of Copyright – Open Future
MEP Axel Voss questions his own copyright framework while offering insights on remuneration and transparency for AI.
openfuture.eu
📄 New White Paper: "Beyond AI & Copyright: Funding a Sustainable Information Ecosystem." @paulk.bsky.social examines how AI impacts information control, argues for public AI models, and proposes a levy system on commercial AI services to support the digital commons. openfuture.eu/publication/...
Beyond AI and copyright – Open Future
This White Paper calls for a levy on commercial AI systems to fund public infrastructures and ensure a sustainable Digital Knowledge Commons in the generative AI era.
openfuture.eu
This destroyed so much more than just that. in the years before the ministry experimented funding for social media initiatives(as in media produced by civil society). this initiative (de digitale pioniersregeling was also abruptly killed.
Agustin Ferrari Braun@ferraribraun.bsky.social · last yr.My niche contribution to the ongoing Platerkerk discourse is that, in 2009, the Dutch media ecosystem was under immense pressure due to the Global Financial Crisis and the rise of digital platforms. The Brikman Commission found that the sector needed 12 million euros per years to survive...
