Bill > HB4668

This bill, titled the "Artificial Intelligence Safety and Security Transparency Act," establishes comprehensive regulations for large developers of artificial intelligence (AI) foundation models with the goal of managing critical risks. The bill defines a "large developer" as an entity that has developed AI models requiring significant computing power (over $5 million for a single model and over $100 million total in the past 12 months), and mandates that such developers create and publish detailed safety and security protocols addressing potential high-stakes risks like the creation of weapons, cyberattacks, or AI-assisted criminal activities. Starting January 1, 2026, these developers must produce a publicly accessible protocol that outlines how they will identify, assess, and mitigate potential critical risks, conduct quarterly transparency reports, and undergo annual third-party audits. The bill also includes robust whistleblower protections, allowing employees to anonymously report potential critical risks without fear of workplace retaliation, and empowers the state attorney general to bring civil actions and impose fines up to $1 million for violations. Importantly, the legislation recognizes the potential societal impacts of advanced AI systems and seeks to establish a proactive framework for responsible development and deployment, with a focus on preventing scenarios that could cause significant harm to human life or property.

Government Affairs, Justice

https://legislature.mi.gov/Bills/Bill?ObjectName=2025-HB-4668