Helping policymakers weigh the benefits of open source AI

Policymakers are increasingly scrutinizing the software components of AI systems, particularly focusing on the availability of AI model weights for downstream use. GitHub emphasizes the importance of understanding developer needs when crafting AI regulations and advocates for AI governance that fosters responsible, secure, and effective development to advance human progress. In response to the U.S. NTIA’s request for comment, GitHub highlighted the benefits of open source AI, which allows developers broad access to weights, code, and other components, thus enhancing innovation, competition, and the global proliferation of AI technology. GitHub argues that regulation should focus on entire AI systems rather than subcomponents, as restrictions on models may hinder beneficial use more than prevent misuse. They stress that government restrictions on AI models are currently unsupported by evidence and recommend prioritizing regulation of high-risk AI systems while preparing for potential abuses. GitHub believes that societal resilience can be achieved not through secrecy but by leveraging the open availability and diversity of AI models, with governments playing a crucial role in advancing AI measurement science, safety research, and public education to foster an AI-driven future aligned with societal values.