Representatives from four leading artificial intelligence companies declined to provide absolute guarantees that their systems would always follow safety protocols during a New York City Council hearing on Monday.

Council Speaker Julie Menin pressed the executives to assure the public under oath that AI agents would comply with guardrails meant to prevent catastrophic harm. Officials from OpenAI, Anthropic, Meta and Google each stopped short of making such a commitment.

OpenAI's Morgan Dwyer stated that the company prioritizes safety and continues to monitor safeguards after deployment. However, Dwyer added that it is not possible to commit or guarantee that any technology is without risk. Anthropic's Logan Graham noted that the science behind controlling these systems remains uncertain.

Google AI and Emerging Tech Policy Director Alice Friend said the company's goal is reliability and safety but emphasized that promising perfection would not be possible with any product on the market. Menin responded that the council was not asking for perfection.

The hearing comes weeks after President Trump announced a voluntary White House Accord on Super Intelligence with executives from OpenAI, Anthropic, Meta, Google, xAI and Nvidia. The one-page agreement calls for internal controls, external audits and board oversight but carries no legal penalties and has been described as morally binding rather than enforceable.

Industry observers have noted that prior voluntary safety commitments by AI firms have seen mixed follow-through in past years. The companies involved in Monday's testimony continue to advance their models while acknowledging inherent limitations in eliminating all risks.

New York City Council members indicated they would continue examining AI oversight measures in future sessions.