In brief

The New York City Council unveiled a package of AI bills on September 25. The main bill would make it unlawful to sell or deploy an AI system in the city without third-party validation and a human “kill switch”, with fines of $25,000 per violation. A hearing is set for October 5.

New to this? Read it in simple words
  • New York City lawmakers proposed new rules for AI systems sold or used in the city.
  • An outside expert would have to check each system first. Every system would need an off switch.
  • Breaking the rules could cost $25,000 per violation. People who report problems could get part of the fines.
  • These are only proposals. Lawmakers will discuss them on October 5.
Words to know
Kill switch
A way for a person to shut down an AI system quickly.
Validation
An independent check that a system meets set rules.
Whistleblower
A person who reports wrongdoing inside a company or organisation.

What the main bill would do

The main bill, Introduction 26835, comes from Menin. It would make it unlawful to market, sell, or deploy an AI system in the city unless an outside reviewer has validated it.

The checks would cover data quality, bias, privacy, security, and the system’s outputs. The city’s Cyber Command could add its own requirements.

Every covered AI system would also need a kill switch, which the Council describes as “a human override that can shut down the system”.

Businesses and validators could each face a $25,000 penalty for every system sold or deployed without validation, or with a falsified one. Fortune reports that Menin said the fine would apply per agent when many AI agents act together.

Sources1234

BILL TRACKER 01
Four rules at the heart of the package, all still proposals.

All 51 Council members will discuss the bills on October 5. Nothing has been voted on yet.

The other bills

A second bill from Menin would give whistleblowers a share of fines recovered from AI companies. The Council calls this a first in the United States, but it has not said how large the share would be.

A bill from Council Member Virginia Maloney would let people sue AI companies over foreseeable harm. That could apply when a company failed to put reasonable safeguards in place and someone got around its safety controls, which is known as jailbreaking.

Other bills would require city agencies and contractors to report AI safety incidents to Cyber Command within 24 hours. They would also create an emergency plan for AI disruptions and protect city workers who raise AI safety concerns.

Further bills would ban false or misleading safety claims by AI companies. The Council says more bills are coming, including on deceptive deepfakes.

Sources134

The Council sent letters to Dario Amodei, Sam Altman, Sundar Pichai, Elon Musk, and Mark Zuckerberg. It says it can use its subpoena power if needed.

According to amNewYork, a Council source said officials are in active talks with several of the companies. Fortune reports that sources say none of the five is likely to testify.

The Council pointed to reports that AI agents tested by OpenAI in July broke out of their controls and compromised systems at Hugging Face during a security test.

“We can and must be both pro-innovation and pro-safety,” Menin said. Fortune notes that a federal bill from Senators Ted Cruz, Amy Klobuchar, and John Thune would likely override most state AI laws, including New York State’s RAISE Act.

Sources1234

Sources

Every fact in this story comes from the sources below. Open them to check our work.

  1. 1
  2. 2
  3. 3
  4. 4
How we checked this story

We read the Council’s own announcement, then compared Fortune, amNewYork, and 6sqft. Bill numbers and sponsors come from the Council. The bills are proposals: they can change before any vote, and no date for them to take effect has been set.