Australia tightens AI safety rules after OpenAI security breach

New York is set to enforce a stringent AI safety law, with large frontier AI developers such as Anthropic and OpenAI required to register by January 1, 2027. This new regulation, part of the state's RAISE Act, mandates these companies to publish safety frameworks on their websites, report critical safety incidents within 72 hours, and provide public model transparency reports.

The Office of Digital Innovation, Governance, Integrity, and Trust (DIGIT) will oversee these requirements. Violations could result in penalties up to $1 million for a single offense and $3 million for subsequent ones. Companies that fail to register may face enforcement actions, though specifics were not disclosed.

Governor Kathy Hochul supports the law, stating that technological progress cannot be ignored, even if there are perceived risks. To back this initiative, Hochul launched the New York Center for AI Responsibility and Research at State University of New York's Binghamton University. This center will conduct independent evaluations of AI systems' safety, security, and operation and work alongside DIGIT to review companies' safety frameworks. The initiative was announced in January, backed by $55 million in donations.

New York is not the only state taking action on AI safety. Last year, the state enacted a law requiring AI companion operators to inform users they are speaking to a bot and to attempt to detect suicidal ideation or self-harm. A separate chatbot safety law for minors, taking effect in June 2027, restricts behaviors like deceptive interactions or manipulation.

Other states like California and Illinois have also implemented similar AI safety and transparency bills, mandating safety frameworks, incident reporting, and independent audits.