New York Legislates to Mitigate AI-Induced Catastrophes

New York State Passes AI Safety Bill Aimed at Preventing Disaster Scenarios

The New York state lawmakers recently passed the RAISE Act, a bill targeting frontier AI models from companies like OpenAI, Google, and Anthropic to prevent disaster scenarios. The bill aims to avoid situations resulting in death or injury to more than 100 people or damages exceeding $1 billion.

Championed by prominent figures in the AI safety movement such as Geoffrey Hinton and Yoshua Bengio, the RAISE Act would establish transparency standards for frontier AI labs, a first of its kind in America. The bill, designed by New York State Senator Andrew Gounardes, sets out to regulate large AI companies without stifling innovation among startups and academia, a criticism faced by California’s SB 1047.

Next Steps for the RAISE Act

The RAISE Act is now awaiting approval from New York Governor Kathy Hochul. If signed into law, it would require major AI labs to publish safety and security reports for their frontier AI models. Failure to comply could lead to civil penalties of up to $30 million, as enforced by New York’s attorney general. The bill applies to companies utilizing over $100 million in computing resources for their AI models and catering to New York residents.

Political Pettiness vs Economic Reasons for AI Companies in New York

Assemblymember Bores expressed his confidence that there is no economic justification for AI companies to withhold their models in New York. Despite acknowledging the potential for political pettiness, he believes that the economic benefits should ultimately prevail.