Open-Source Coalition Pushes California to Rework AI Act
Data Breach TodayArchived Jun 25, 2026✓ Full text saved
Developers Warn Clause in AI Transparency Act Collides With Open-Source Licensing A coalition of open-source artificial intelligence players are pressing California to rewrite a license-revocation provision in the state's AI Transparency Act, warning that the language as drafted clashes with how open-source licensing works and could seed uncertainty across the software supply chain.
Full text archived locally
✦ AI Summary· Claude Sonnet
Artificial Intelligence & Machine Learning , Legislation , Next-Generation Technologies & Secure Development
Open-Source Coalition Pushes California to Rework AI Act
Developers Warn Clause in AI Transparency Act Collides With Open-Source Licensing
Chris Riotta (@chrisriotta) • June 24, 2026
Credit Eligible
Get Permission
Image: Shutterstock/ISMG
GitHub and a coalition of open-source artificial intelligence players are pressing California lawmakers to overhaul a provision in state law meant to make it easier for users to identify content generated by AI, saying it clashes with how open-source licensing fundamentally works.
See Also: Securing AI Workloads With Ubuntu Pro
The coalition - GitHub, Black Forest Labs, Hugging Face and Mozilla - called on legislators to rewrite a license-revocation provision in the bill, writing in a letter to state Sen. Josh Becker that it could lead to uncertainty across the software supply chain.
Under the California AI Transparency Act a company that builds a generative AI system has 96 hours to revoke a license after discovering a third party has altered the system so it no longer attaches the disclosures the law requires. But, a fundamental tenant of open source is that licenses are irrevocable, a measure meant to safeguard the openness of open source.
"If licenses were revocable at the discretion of an upstream developer, that would destabilize or disrupt entire downstream systems," the letter argues.
The California law is a first-of-its-kind bill that Gov. Gavin Newsom signed in 2024, making the state the first to mandate built-in provenance disclosures and detection tooling for AI-generated content. It provisions begin staggering into effect starting Aug. 2. The law applies to companies that create, code or otherwise produce a GenAI system with more than 1 million monthly users and publicly accessible in California.
"This law creates material harms to open-source licensing" and does not resolve "the fundamental incompatibility between the proposed license revocation requirements and open-source licensing," the letter reads.
"There is a better way forward that advances the bill's objectives and maintains compatibility with open-source development," the letter continues, pointing to the EU AI Act's Code of Practice on Transparency. The code of practice requires covered providers to make "best efforts" to maintain the integrity of a provenance system while stipulating that providers meet the requirements by notifying customers of their legal obligations, rather than fully revoking their licenses.
Developers who modify and deploy AI systems are already directly covered by the statute, the letter says, and its enforcement mechanisms remain in place without it.
The coalition's fix targets SB 1000, a measure from Becker that would amend the act that cleared the Senate on a 33-1 vote before heading to the Assembly.
Those providers must offer a free public detection tool, give users the option of a visible disclosure and embed a hidden provenance disclosure in AI-generated image, video and audio content. Text outputs fall outside the scope of the bill.
Violations can receive civil penalties of $5,000 per violation, with further penalties available for licensees that keep running a revoked system.
California lawmakers have argued the bill is required to help combat misinformation and disinformation tied to public health, election integrity, fraud and extortion through GenAI.