NYC Grills AI Executives, but Whistleblowers Say Speed Still Wins
NYC Grills AI Executives, but Whistleblowers Say Speed Still Wins
The confrontation had been building for weeks. As scrutiny of advanced AI intensified, City Council Speaker Julie Menin called a rare Committee of the Whole hearing, seeking sworn testimony from Anthropic, OpenAI, Google and Meta on threats posed by their systems and possible legislation. The companies agreed to appear only after the council raised the prospect of subpoenas, according to Menin’s office.
On Monday, the hearing exposed the gulf between the council’s demand for specifics and the industry’s broad assurances. Menin asked representatives how they quantified the risk in a worst-case catastrophic scenario. OpenAI’s Morgan Dwyer replied, “I don’t know,” adding that she did not believe the difference between a 1%, 10% or 20% chance of catastrophe mattered because no such level would be acceptable. She said OpenAI should not train models that cannot remain safely under human control.
That answer did not satisfy the council. Menin called it “flippant at best,” while Public Advocate Jumaane D. Williams branded it a “God-awful, fantastically terrible answer.” The dispute went beyond rhetoric: lawmakers were asking whether companies could halt a model that failed safety reviews, insure against catastrophe, or even put a number on the danger. They left with little clarity.
Former workers then supplied the hearing’s sharpest counterpoint. Jacob Coxon, who resigned from Anthropic, said the familiar Silicon Valley instinct to “move fast, break things” may suit a photo-sharing app, but “does not work for the most powerful technology ever built.” He appeared with former Google and OpenAI employees warning that commercial pressure was outrunning safeguards.
Their message aligned with a broader push for intervention: AI insiders urged greater oversight as local governments begin confronting the possibility that rapidly advancing systems could endanger humanity. The hearing left a stark split in view—companies promising intolerance for catastrophe, and critics asking why they cannot yet demonstrate how they would prevent it.
Write a comment