Recommended Free Tools
The New York City Council held a citywide hearing on October 5, 2026, to question technology companies about AI risks, company safeguards and possible local rules. Anthropic, OpenAI, Google and Meta were announced as public witnesses; the Council also said it subpoenaed SpaceXAI after it did not respond to an inquiry. The proposals discussed are not established here as enacted laws, so they should be read as potential requirements—not rules already in force.
What the Council hearing covered
The Council’s Committee of the Whole convened all 51 members, according to the Council’s September 28 announcement. It said the hearing would examine risks to New Yorkers from rapidly developing AI, safeguards used by major companies, and possible city legislation or regulation. The Council described it as a rare Council-wide hearing.
The Council said Meta had confirmed a senior leader’s appearance earlier. OpenAI and Google agreed to testify after the Council warned of possible subpoenas, and Anthropic agreed shortly before a subpoena was due. The Council said SpaceXAI had not responded to its inquiry and that Speaker Julie Menin issued a subpoena compelling it to appear. The announcement alone does not establish that SpaceXAI attended or complied.
Public Advocate Jumaane D. Williams also questioned representatives of OpenAI, Anthropic, Meta and Google and delivered prepared remarks, according to his office’s October 5 statement.
#1 Best Overall
What witnesses said about AI risks and safeguards
Associated Press reporting on the hearing described contrasting testimony: former employees of Anthropic, OpenAI and Google DeepMind raised concerns about company safety culture and control risks, while representatives of OpenAI, Anthropic, Google and Meta defended their focus on safety and described benefits of their systems. These are positions expressed by witnesses, not findings that establish the systems’ safety or the likelihood of a particular outcome.
Former employee Jacob Coxon’s warning
Former Anthropic engineer Jacob Coxon criticized what he characterized as a rapid-development culture: “The companies run on a startup mindset: Move fast, break things, fix them later. That works for a photo sharing app. It does not work for building the most powerful technology ever built,” he told the Council, as reported by the Associated Press on October 5.
Rank #2
Coxon also said, “On the current path, I think it is more likely than not that humanity loses control to these AIs and it could end in human extinction.” That is his personal warning; it is not a measured probability or an established consensus. AP reported that Speaker Menin asked company representatives to quantify catastrophic risk, but the report does not provide a complete, verified response from every representative.
Company and Public Advocate perspectives
OpenAI representative Morgan Dwyer told the Council, “I share your commitment to ensuring that technology is developed safely and that its benefits are widely shared,” according to AP. The report described company representatives as emphasizing safety and benefits and expressing openness to some Council suggestions.
Williams urged a precautionary approach in his prepared remarks: “Rather than waiting to uncover harm, we should be proactive, limiting exposure until it has been shown to be safe and the promised benefits have materialized.” He also asked company representatives, “What is the guardrail, that if we put it up, you would push back the hardest?” Those statements reflect the Public Advocate’s position and question, not a Council finding.
What AI rules the Council proposed
The Council’s September 25 announcement described proposals scheduled for consideration at the October 5 hearing. The measures below are proposal terms as summarized by the Council; the materials cited here do not establish each bill’s later legislative status.
Duties for AI providers and chatbot operators
- Introduction 2602 — validation and human override: Sponsored by Speaker Julie Menin, this proposal would require third-party validation of covered AI systems marketed, offered for sale or deployed in the city. The Council said validation would address data quality, bias, outputs, privacy and security. The proposal also includes a human override or “kill switch,” validator conflict disclosures and a proposed $25,000 penalty per instance for certain violations.
- Introduction 2603 — company disclosures: This proposal would require disclosures about AI tools and prohibit false or misleading safety claims.
- Introduction 2599 — chatbot safeguards: This proposal would set local privacy, security and transparency requirements for chatbot providers.
City response, reporting and remedies
- Introduction 2601 — safety incident reporting: Contractors or agencies would have to report covered AI safety incidents to the Office of Cyber Command within 24 hours. The office would have to disclose them publicly within 24 hours, according to the Council’s summary.
- Introduction 2606 — emergency planning: Cyber Command and NYC Emergency Management would be required to plan for AI-related events that compromise city systems or infrastructure, or disrupt government operations or public safety.
- Introduction 2600 — private right of action: This proposal would create a private claim for certain foreseeable harms arising from malicious AI use or circumvention of safety controls. The Council’s description includes conditions concerning foreseeability, failure to implement reasonable safeguards and exploitation of that failure by a third party.
- Introduction 2605 — whistleblower incentives: Individual whistleblowers could receive a portion of fines or penalties recovered from AI companies for applicable violations.
- Introduction 2604 — employee and contractor protections: This proposal would clarify protections for certain city employees and contractors who report AI-related conduct they reasonably believe presents a public safety threat.
Workers, elected officials and manipulated media
- Introduction 161 — effects on city workers: This proposal would expand reporting on how algorithmic tools affect city employees and employment responsibilities.
- Introduction 504 — likenesses of elected officials and candidates: Elected officials and candidates could give written notice that they do not authorize systems to generate manipulated media depicting their likeness. The Council’s release describes restrictions and penalties in the proposal.
- The Council also said additional proposals would address deceptive deepfakes and the effects of algorithmic tools on job duties.
How the proposals differ
| Who or what is affected | Examples of proposed mechanisms |
|---|---|
| AI developers and providers | Third-party validation, human override, disclosures, and chatbot privacy, security and transparency requirements. |
| City agencies and contractors | Safety incident reporting and planning for disruptions to city systems, government operations and public safety. |
| People seeking remedies or reporting violations | A private claim for certain harms, whistleblower shares of recovered penalties, and protections for certain city workers and contractors. |
| Workers and people whose likenesses may be manipulated | Expanded reporting on workplace effects and a notice mechanism for elected officials and candidates concerning manipulated likenesses. |
The Council also said it enacted AI-related city laws in 2025 concerning an Office of Algorithmic Accountability, safeguards for city agency development and procurement, and a public inventory of AI systems subject to review. Those earlier laws are separate from the 2026 proposals; the Council’s announcement does not establish their current implementation details.
What is—and is not—established about the rules
The hearing and the Council’s September announcements establish that the Council put forward proposals and heard testimony. They do not establish that every proposal passed, was signed into law or is currently enforceable. Before relying on any measure as a legal obligation, check the Council’s current legislative record for its status and final text.
Best Value
Likewise, testimony about catastrophic risks should not be mistaken for statistical evidence. The materials cited here provide no independently established statistical estimate of catastrophic AI risk; Coxon’s “more likely than not” statement is his attributed opinion. The proposed $25,000 penalty and 24-hour reporting and disclosure windows are terms the Council summarized for specific proposals, not evidence that the measures took effect.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
