AI developers must disclose training data sources

File S06955 ("AI developers must disclose training data sources") is before the New York Legislature and has not been decided. Latest step (March 4, 2026): ADVANCED TO THIRD READING.

New York file S06955, before the New York Legislature. Still to be decided as of 2026-03-04.

What it does

See what data trained AI you rely on When you use a generative AI tool (ChatGPT, image generators, etc.), the developer must post a public summary of the datasets used to train it on their website.

The bigger picture

May shift AI company practices and transparency norms Disclosure requirements could pressure AI developers to audit and document training data more carefully, potentially reducing use of unlicensed or scraped content and encouraging industry-wide transparency standards.

Read the official record for S06955

See everything up for a vote · This week in New York · See what we have filed with the clerk