Industry · New York State ·
New York’s RAISE Act takes effect 1 January 2027: what frontier AI labs must do
Large AI model developers face annual safety reviews, third-party audits and 72-hour incident reporting.
What happened
New York’s Responsible AI Safety and Education (RAISE) Act applies from 1 January 2027 to frontier model developers deploying in New York whose training runs exceeded 10²⁶ operations or cost more than $100 million.
Requirements include annual safety reviews, independent third-party audits, reporting incidents within 72 hours, and a ban on deploying models that pose an “unreasonable risk of critical harm”.
What it means for you
This targets the big labs, not app developers — but expect stricter usage policies from model providers as they comply.
Sources
Written by TheEvents from public reporting and the companies’ own announcements. Facts can change as stories develop — check the sources for the latest.
Related
Jensen Huang resigns as Nvidia CEO; Colette Kress named interim CEO
Nvidia’s founder left the CEO, president and board roles. CFO Colette Kress takes over on an interim basis.
Fractile raises ~$720M at a $7B valuation for inference chips
The London chip startup puts memory next to compute for faster, cheaper inference, and has a $250M supply deal with Anthropic.