GitHub Copilot's use of GPL code for training is not infringement; adopting a maximalist copyright stance will ultimately hurt the Free Software community
Reda's position sits against an earlier EU proposal that would have required code-sharing platforms to monitor uploaded material for copyright infringement, making expansive enforcement rules consequential for repositories as well as individual developers.
The subsequent arc moved from principle to product and litigation: GitHub added corporate licensing and policy controls for Copilot, while Microsoft, GitHub, and OpenAI later defended training practices on fair-use grounds.
First-order effects
Reda's intervention gives GPL and free-software advocates a case against treating Copilot's training use as infringement, rather than a call for broader repository monitoring.
GitHub and Copilot face a sharper governance debate over open-source code use, even though Reda's view does not itself change the product's legal status.
Second-order effects
Corporate buyers’ demand for controls becomes a practical response to the uncertainty around generated code, reflected in Copilot for Business's licensing and policy features.
The dispute shifts pressure toward courts to define the limits of training on open-source code; the later developer suit narrowed substantially when a judge dismissed most of the original claims.
Third-order effects
If the fair-use position continues to prevail in court, disputes over AI coding tools are likely to concentrate less on access to public code and more on output attribution, licensing controls, and enterprise risk management.
The broader structural tension is between open-source communities seeking enforceable reciprocity and platforms seeking to train and commercialize tools from publicly available code.
The trend: AI coding assistants are turning open-source licensing from a repository-level obligation into a contest over training rights, generated outputs, and commercial controls.
github copilot has, by their own admission, been trained on mountains of gpl code, so i'm unclear on how it's not a form of laundering open source code into commercial works. the handwave of “it usually doesn't reproduce exact chunks” is not very satisfying https://twitter.com/..…
Hi. I know you're excited about copilot. GitHub scraped your code. And they plan to charge you for copilot after you help train it further. It's truly disappointing to watch people cheer at having their work and time exploited by a company worth billions.
“The output of a machine simply does not qualify for copyright protection - it is in the public domain. That is good news for the open movement and not something that needs fixing.” https://twitter.com/...
“I'm leaving GitHub because copilot uses my OpenSource code for training” is such an odd move. Anyone can fork it to there and GitHub can feed OpenSource code from anywhere to it and US copyright law permits this. I'm also pretty certain we should not strengthen copyright laws ..…
Some thoughts about Github Copilot and copyright (I am not a lawyer). Copilot uses a version of GPT3 trained on GPL licensed code. GPL give everyone the right to copy and make derivatives. Derivatives inherit GPL. Copilot can sometimes memorize and repeat snippets of code. https:…
When it comes to digital policy, @Senficon is nearly always right. Placing AI-constructed works under copyright is a very bad idea. https://juliareda.eu/...
Neural weights are not code themselves and do not resemble the surface form of any GPL code. GPL covers the possibility of “non-source conveyance”. This would apply if neural weights are deemed a form of “object code”. There are reasons to reject this line of argument, tho.
Copilot can't run without the weights of the neural net and those weights cannot exist without the data. Is the model part of the “executable code” of Copilot? There are reasons why the answer could be yes, in which case Copilot would be GPL...
i'm really tired of the tech industry treating neural networks like magic black boxes that spit out something completely novel, and taking free software for granted while paying out $150k salaries for writing ad delivery systems. the two have finally fused and it sucks
copyright does not only cover copying and pasting; it covers derivative works. github copilot was trained on open source code and the sum total of everything it knows was drawn from that code. there is no possible interpretation of “derivative” that does not include this
Free Software advocates believed the GPL would end companies making money from proprietary software. Instead it simply enabled new types of proprietary software companies like Google & Amazon. Github Copilot is continuing that tradition. Market incentives beat simple solutions. h…
Open Source means everyone can see, share, and use your code even if you disagree with them. If you want to pick and choose who can use your code, don't fucking Open Source it. https://twitter.com/...
More and more, well-established open-source developers are abandoning Github. This time it is about ML-powered GitHub Copilot product. https://thelig.ht/... What do you think? https://twitter.com/...
Increasing number of developers are upset Github CoPilot is trained on Open Source code and then used to build a proprietary product. Regardless of legality, this is similar to https://clearview.ai/ which was trained on public photos. Legal yet creepy. https://thelig.ht/... https…