New York holds hearing on AI regulation as Brazil debates technology’s use in elections

Representatives from leading AI companies—OpenAI, Anthropic, Google, and Meta—appeared at a sworn hearing in the New York City Council on Monday, June 5. The session, convened by Council President Julie Menin, brought together council members, experts, and former industry insiders to discuss oversight and accountability measures in response to AI system failures and abuses.

The companies participated under pressure, with OpenAI, Google, and Anthropic accepting the invitation only after facing the threat of a court subpoena. Meta had already assigned an executive to attend the meeting.

Independent audit and shutdown button

A proposed bill would ban the sale and deployment of systems in the city that haven’t undergone independent technical validation. This assessment would cover data quality, security, privacy, and the presence of biases.

The same proposal requires a mechanism that allows a human operator to disable the AI. Noncompliance could result in a $25,000 fine for each deployment carried out without validation or with a falsified assessment.

Another initiative would allow individuals harmed by the malicious use of AI to sue companies for foreseeable damages if the companies failed to implement reasonable safeguards.

Council members also discussed incentives for whistleblowers and the requirement to notify the municipal cybersecurity agency within 24 hours of an incident.

Warnings from former employees and security concerns

Former employees of AI labs participated in the hearing and warned about the rapid pace of development, according to the Associated Press. Their testimonies followed incidents in which AI agents gained access to real systems during testing, exposing flaws in the containment environments meant to limit the models’ actions.

Brazil debates AI, disinformation, and elections

On June 1, the Social Communication Council of Congress held a hearing on disinformation and the use of artificial intelligence in the 2026 elections. Researchers highlighted challenges in monitoring digital advertising and called for greater transparency.

While Brazil’s discussion focused on informational and electoral risks, New York’s proposals also address technical validation, the ability to shut down systems, and liability for damages. A key question in both countries is what obligations should be imposed on companies before and after an incident occurs.