AI and the Destruction of the Creative Commons

Sharing your code online now risks it being absorbed into AI models without compensation or adherence to your original license.

Generative AI large language models (LLMs) are fundamentally disrupting the established norms of software copyright and open-source licensing. Historically, open-source licenses like GPL and CC-SA ensured that derivative works inherited the same freedoms as the original code. However, LLMs are trained on vast datasets scraped from the internet without explicit regard for these licenses. This process raises concerns that derivative works created with AI may not adhere to the original creators' licensing terms, effectively breaking the social contract. The author argues that this trend disincentivizes sharing code and other creative works, as they can be absorbed and exploited without proper attribution or adherence to original freedoms. Furthermore, open-source code may become more vulnerable to exploitation and the influx of AI-generated noise, making public sharing a liability rather than a benefit. The article posits that this represents a move toward a 'digital dark age,' undoing decades of progress driven by openness.

The foundational principles of open-source software and creative commons licensing, which fostered decades of online innovation and collaboration, are under severe threat. By consuming and processing vast amounts of licensed data without respecting original terms, generative AI models risk creating a new generation of derivative works that are disconnected from their source freedoms. This erodes the legal and ethical framework that empowered creators and fueled the modern internet, potentially leading to a future where innovation is stifled by a lack of trust and incentive to share.

The implications extend beyond copyright, impacting the very nature of collaboration and knowledge sharing. Creators now face increased risks of their work being exploited, their code being used to identify vulnerabilities, or their public repositories being overwhelmed by AI-generated submissions. This shift transforms open sharing from an act of empowerment into a potential liability, concentrating wealth and power with AI developers rather than the original creators and signaling a potentially damaging reversal of decades of open digital progress.