Summary#
Introduction 1130-2026, sponsored by Council Member Carl Wilson, would amend the New York City Administrative Code by adding rules for advertisements promoting artificial intelligence models. Its stated goal is to require disclosures about third-party validation and prohibit materially false or misleading claims about certain risks, risk-management measures, and validation. It is not law; the supplied status says it is in committee.
- A person who advertises an AI model in the city would have to disclose whether it was validated by a third-party validator under an existing Administrative Code subchapter. The commissioner must make rules about the disclosure’s form.
- Advertisements could not make materially false or misleading statements about a model’s “substantial risk,” measures taken or planned to manage that risk, or whether the model was validated.
- A violation could bring a civil penalty of up to $25,000 per violation. The Corporation Counsel may sue to recover penalties and seek an order requiring or stopping action, legal costs, and other remedies.
- The law would take effect 180 days after it becomes law.
What it means for you#
- Businesses and people promoting AI models: If you disseminate, or cause someone to disseminate, a commercial message in any medium promoting an AI model in New York City, you would have to disclose whether the model was validated. The bill does not require that the model be validated; it requires an accurate disclosure.
- Advertisers: You could not make materially false or misleading claims about the model’s defined substantial risks, steps taken or planned to manage those risks, or its validation status. The bill defines substantial risk to include foreseeable and material risks of death or serious injury, property damage or loss, or certain breaches of security resulting from specified model conduct.
- Consumers and the public: The bill would regulate advertising, not create a general right to have AI models independently tested or a new process for people to report ads.
- City government: The commissioner must set the disclosure’s form. The Corporation Counsel may bring a civil case, but the bill does not require the office to sue in every case.
Money#
No cost information is in the available material.
- The bill requires the commissioner to make rules, but gives no estimate of the cost or staffing needed.
- A person who violates the rules could face a civil penalty of up to $25,000 per violation. If a deceptive ad also violates an existing city law, the bill allows a separate penalty or fine under that law as well.
What is unclear#
- The bill refers to third-party validation under another Administrative Code subchapter, but that subchapter was not supplied. I could not verify the existing validation rules or how this bill would interact with them.
- The bill does not name the commissioner who must set the disclosure’s form.
- It does not specify when or how the commissioner must issue the rules, or how violations would be identified and handled before a lawsuit.
- “Materially false or misleading” is not defined in the bill. The text also does not explain how its requirements interact with the existing city provisions it cites.
Case for#
- A possible argument for the bill is that people seeing AI advertisements would learn whether the promoted model underwent third-party validation.
- Requiring accurate claims about substantial risks and risk-management measures could make it harder to mislead the public about those issues.
- The bill identifies a city enforcement route and allows a court to order a person to stop or take action.
Case against#
- One concern is that the bill does not set the disclosure’s form or timing; those details would depend on later rules.
- The bill’s reference to existing validation rules makes the effect harder to assess without those rules. It also does not explain how the cited city laws would apply alongside the new penalties.
- The bill allows penalties of up to $25,000 per violation but does not specify in this text how a violation would be counted.