Requiring reporting and public disclosure of artificial intelligence safety incidents concerning city contracts

Summary#

Int. 1109-2026, sponsored by Council Member Kamillah Hanks, is a proposed local law amending the New York City Administrative Code. Its stated goal is to require reporting and public disclosure of certain AI safety incidents connected to city contracts. It is referred to the Committee of the Whole and remains in committee; it is not law.

  • It covers contracts entered into after the law takes effect when their performance would foreseeably involve developing, storing, using, or deploying an AI model. Reportable incidents include specified harms, security breaches, loss of control, and some attempts by a model to evade its developer’s controls.
  • The Office of Cyber Command (NYC3), consulting the city’s chief procurement officer and the Office of Technology and Innovation, must set and periodically review contractor procedures for identifying incidents. Covered contracts must require contractors to report incidents to NYC3 in writing within 24 hours of becoming aware; agencies have the same deadline.
  • NYC3 must post a description, date, and reasons for each incident within 24 hours of becoming aware, subject to limits for trade secrets, cybersecurity, public safety, national security, and information barred from disclosure by law. It must also publish an annual report with details including the contractor, agency, and how each incident was resolved. The law would take effect 180 days after it becomes law.

What it means for you#

  • City contractors: If a covered contract includes AI work, the contractor must follow NYC3’s procedures and report qualifying incidents within 24 hours of learning of them.
  • City agencies and NYC3: Agencies must report incidents they learn about. NYC3 must set procedures, disclose incidents, and issue annual reports.
  • Public: The bill requires online disclosure, but allows specified information to be withheld. It does not establish fines or require the city to ban AI use.

Money#

No cost estimate is provided in the available material.

  • The bill requires NYC3 to create and maintain procedures and publish disclosures and annual reports, but does not specify staffing or funding.
  • It sets no fee or fine.

What is unclear#

  • The current Administrative Code was not supplied, so existing reporting duties or overlap with them could not be verified.
  • The bill uses terms such as “materially increased substantial risk” and “otherwise relates to any contract”; how those apply in particular cases may require judgment.
  • The bill does not specify enforcement for missed reports or how NYC3 will handle disputes over whether an incident qualifies or what information can be withheld.

Case for#

  • A possible argument for the bill is that quick reporting and public disclosure could help the city and the public learn about serious AI-related harms or security breaches connected to city contracts.
  • The bill requires NYC3 to set procedures, which could give contractors a common method for identifying incidents.

Case against#

  • One concern is that the 24-hour deadlines may be difficult to meet when the incident’s cause or significance is not yet clear; the bill does not describe a process for resolving disagreements.
  • The bill requires public disclosure but does not explain how NYC3 will balance timely notice against the listed reasons for withholding information.