NYC Council Calls AI Giants to Testify on Control Risks in Landmark Hearing

On October 5, 2026, the New York City Council convened a high‑profile, sworn hearing that brought senior executives from four of the world’s largest artificial‑intelligence laboratories—Anthropic, OpenAI, Google and Meta—into a formal session before city lawmakers. The hearing was explicitly framed around the growing concern that advanced AI systems could act in ways that diverge from human values, and it highlighted the legal uncertainty that still surrounds accountability when such systems cause harm.

The panel of witnesses featured Logan Graham representing Anthropic, Morgan Dwyer speaking for OpenAI, Alice Friend on behalf of Google, and Shane Cahill as Meta’s senior liaison. Their participation came after the council threatened to issue subpoenas; Meta had already agreed to appear, while the other three firms only consented after the council announced the possibility of legal enforcement.

Key concerns raised by experts

Former DeepMind researcher Alex Turner warned that AI systems might pursue objectives that are fundamentally misaligned with human intent, creating a risk of catastrophic outcomes if those systems are left unchecked. Turner’s comments echoed the longstanding warnings from AI‑safety scholars about the danger of ‘goal misalignment’ in increasingly autonomous models that operate with limited human oversight.

Two former employees—Jacob Coxon, who previously worked at Anthropic, and Daniel Kokotajlo, a former OpenAI researcher—called for greater transparency regarding development roadmaps. They urged a deliberate slowdown of work that pushes the performance frontier, emphasizing an internal debate within the industry about balancing rapid innovation against robust safety safeguards.

Council’s critique of risk quantification

Council President Julie Menin criticized the testimony for lacking a clear, quantitative assessment of catastrophic risk. She argued that without concrete metrics, policymakers cannot design effective mitigation strategies or allocate resources in a way that meaningfully reduces the danger posed by advanced AI.

Menin also highlighted the legal gray area surrounding liability. “If an AI system causes damage, who is responsible?” she asked, underscoring the urgent need for a legislative framework that defines accountability for both developers and deployers of powerful AI technologies.

Industry positions on regulation

Google’s Alice Friend defended a federal‑level, uniform regulatory approach, contending that a patchwork of local rules would stifle innovation and create compliance chaos for multinational firms that operate across many jurisdictions.

In contrast, Anthropic’s Logan Graham acknowledged the council’s concerns but suggested that voluntary industry standards could complement, rather than replace, any future federal legislation, allowing the sector to move more swiftly while still addressing safety gaps.

• OpenAI’s Morgan Dwyer pledged to increase transparency on model capabilities
• Anthropic’s Logan Graham promised to publish a risk‑assessment framework
• Meta’s Shane Cahill committed to a cross‑industry safety working group
• Google’s Alice Friend urged Congress to enact a national AI safety bill

SpaceXAI, a newer entrant in the AI field, failed to appear despite a formal subpoena. The council announced it will pursue a court order to compel the company’s attendance, sending a clear signal that non‑compliance will be met with legal consequences.

For English‑speaking organizations, the hearing marks a turning point. It demonstrates that city‑level authorities are willing to use their subpoena power to enforce testimony, and it suggests that the current lack of a unified regulatory regime may soon give way to comprehensive federal legislation.

The council’s actions also hint at a future where companies operating in the United States should anticipate stricter reporting obligations, mandatory safety audits, and clearer liability rules that could affect product rollout timelines, insurance costs, and overall business strategy.

Legal experts noted that the hearing could set a precedent for other municipalities to follow suit, potentially creating a cascade of local investigations that pressure the industry to adopt higher safety standards even before federal laws are enacted.

Stakeholders from civil‑society groups praised the council for taking a proactive stance, arguing that public oversight is essential to ensure that powerful AI systems are developed responsibly and that the public interest is protected.

Conversely, some industry analysts warned that overly aggressive regulatory demands could drive talent and research investments abroad, where regulatory environments might be perceived as more favorable to rapid AI development.

The hearing concluded with a commitment from the council to draft a set of policy recommendations that will be presented to New York State legislators, with the aim of shaping a cohesive approach to AI safety at both the state and federal levels.

Local officials in Manhattan and Brooklyn expressed particular concern about the deployment of AI‑driven surveillance tools in public spaces, emphasizing the need for clear safeguards to protect civil liberties while still leveraging technological benefits for city services.