GitHub вступил в коалицию против конфликта закона о прозрачности ИИ с open source

GitHub has joined a coalition of companies seeking amendments to California's AI transparency law. The platform insists on targeted changes that will resolve conflicts with open source licenses without undermining regulators' intent.

What's Wrong with the Law

The AI transparency law requires companies that create or deploy AI systems in California to disclose information about model training methods, data used, and decision-making logic. The goal is to enable users and regulators to understand how systems that affect their lives work. The problem is that standard open source licenses — MIT, Apache 2.0, GNU GPL — were created with different logic. They require preserving license text and providing attribution, but say nothing about training data transparency or architectural decisions. A developer publishing an AI model under an MIT license technically complies with its requirements — but may become a California law violator for not providing additional documentation the license does not require.

• Independent researchers publishing models on GitHub risk automatically violating the law
• Non-commercial AI open source projects fall into the same gray zone as corporate developments
• International teams releasing software for California users will face legal uncertainty
• Universities and research institutes will be subject to the same requirements as major technology companies

GitHub's Position in the Dispute

GitHub does not oppose AI regulation as such. The company calls for targeted amendments: clearly define who the law applies to, establish exceptions for open projects, and align California requirements with international standards.

"We call for targeted amendments that will resolve conflicts with open source licenses, align the law with international transparency standards, and preserve regulatory intent," — from

GitHub's official position.

Special emphasis is placed on compatibility with the European AI Act — the most developed international document in this field. It contains explicit exemptions for open models with reduced risk levels. If California establishes stricter and incompatible rules, this will create additional legal burden for companies operating in both markets.

Context: California and AI

California has spent several years attempting to establish an AI regulatory system. In 2024, Governor Newsom rejected SB 1047 — a law requiring companies to test large models for safety before public release. Critics pointed out that the law affected the open community equally with corporations: requirements were identical for Meta with its Llama and for three university researchers.

The transparency law is the next step with a narrower purpose: not to control development, but to ensure visibility for users and regulators. But the problem remains: universal requirements make no distinction between closed commercial products and open projects, where all documentation is already public by definition. GitHub is the central platform for global open source code, including thousands of AI models, datasets, and tools.

Any regulation affecting open source directly concerns its ecosystem.

What This Means

GitHub's participation in the coalition is a signal that major technology platforms intend to actively participate in shaping AI regulation, rather than simply adapt to it after the fact. For open source developers, this is an important precedent: their interests are represented at the legislative table. The outcome of the California discussion will set the tone for other states and likely influence future federal initiatives.

*Meta has been recognized as an extremist organization and is banned in the Russian Federation.