Former employees of Anthropic, OpenAI and Google DeepMind spoke at a New York City Council hearing on Monday.

They warned that the biggest AI companies are racing to build technology they do not know how to control.

Jacob Coxon is a former Anthropic engineer.

He said the firms follow a “move fast, break things” startup mindset.

He said this approach does not suit the most powerful technology ever built.

Coxon said AI could bring great benefits.

But he said there is no transparency and few safeguards.

Because of this, humanity will probably lose control of these systems.

He warned this could end in human extinction.

He resigned and shared his fears.

Days later, top executives at Anthropic, OpenAI and SpaceXAI agreed that AI development needs to slow down.

A second panel of staff from OpenAI, Anthropic, Google and Meta also testified.

They said their systems are already improving daily life.

They also said they focus strongly on safety.

OpenAI’s Morgan Dwyer said he did not know the worst-case risk level.

He said any chance of catastrophe is unacceptable.

Council speaker Julie Menin called that response “flippant at best.”

The Council is considering rules that would require companies to disclose information about AI tools.

Other proposals would require third-party approval for city AI systems.

They would also let whistleblowers share in fines.

New York Gov. Kathy Hochul and other governors have also pushed for reporting and transparency standards.

Anthropic’s Logan Graham said the company welcomes “smart regulation.”

He said state and local governments have a role in building frameworks.