Overview
- The City Council on Sept. 25 introduced a package that would bar marketing, sale or deployment of covered AI systems in the five boroughs unless an outside validator checks data quality, bias, privacy and security and the system includes a human override with $25,000 fines for each uncured instance.
- Council Speaker Julie Menin set a Committee of the Whole hearing for Oct. 5 and invited major AI CEOs to testify, the council has reserved subpoena power and sources say those chief executives are unlikely to appear.
- One bill would pay whistleblowers a share of fines recovered from companies that break the law, and a companion measure would create a private right of action for New Yorkers harmed when foreseeable risks are not guarded against or safety controls are circumvented.
- Other proposals require city contractors to report AI safety incidents within 24 hours, ban false or misleading safety claims, impose privacy and transparency rules for chatbots, and force the city to plan for AI-driven attacks on municipal systems.
- The measures are designed to protect New Yorkers and keep firms in the city but raise unresolved legal and technical questions about federal preemption, Section 230 coverage and the feasibility of independent validation and kill switches that could prompt court challenges.