The RAISE Act (Responsible AI Safety and Education Act) was passed by the New York legislature in 2025. It targets organizations that train the most compute-intensive and expensive frontier models, and it centers on safety protocols, incident disclosure, and transparency rather than everyday AI use.
The threshold for a covered developer is defined by very high training compute and cost, a short list of the largest model builders. If your company uses AI tools built by others, you are almost certainly a deployer. Your exposure is less about building safe frontier models and more about knowing, and proving, which AI your team runs.