On Monday, cognoscenti within the artificial intelligence milieu apprised New York City’s leadership of the perils inherent in this nascent technology, advocating for more stringent oversight as municipal authorities endeavor to address escalating apprehensions that such systems might imperil humanity.

At a hearing convened by the New York City Council, erstwhile employees of Anthropic, OpenAI and Google DeepMind — recently departed from those firms — portrayed a culture of recklessness pervading them, contending that the foremost technology companies are engaged in a race to create technology whose control they do not comprehend.

"The companies operate on a startup ethos: move fast, break things, remediate later.

That may suffice for a photo-sharing application.

It is wholly inadequate for building the most powerful technology ever devised," said Jacob Coxon, a former engineer at Anthropic who has become a leading voice for regulation.

State and municipal governments have been deliberating over mechanisms to regulate AI, given that the Trump administration has adopted a posture of passivity and Congress has demonstrated itself to be inert.

President Donald Trump has dismissed apprehensions surrounding AI as a “hoax” and, late last month, affixed his signature to an agreement delineating a series of voluntary measures for the companies to undertake, one that largely permits them to self-police.

Coxon was a scarcely known engineer at Anthropic until he announced his departure through a series of social media posts that caught global attention.

He told the Council that AI could have tremendous upsides, but absent transparency and safeguards, it’s more likely that things will spiral out of control.

“Given the trajectory we are presently on, I consider it more probable than not that humanity will forfeit control to these AIs, a scenario that could culminate in human extinction,” Coxon said.

Scarcely days after Coxon tendered his resignation and gave voice to his apprehensions regarding worst-case eventualities, a coterie of preeminent executives—among them Dario Amodei of Anthropic, Sam Altman of OpenAI, and Elon Musk of SpaceXAI—professed concordance with the proposition that AI’s development must be decelerated.

AI enterprises have enumerated the measures they have undertaken to accord primacy to safety, yet any concerted industry-wide retardation confronts a multiplicity of impediments.

A coterie of personnel from AI enterprises — representing the tech behemoths OpenAI, Anthropic, Google and Meta — likewise gave testimony at Monday's Manhattan hearing, apprising officials that their systems are already ameliorating people's quotidian existence whilst underscoring that they remain resolutely fixated upon safety.

Anthropic and OpenAI have asserted that their advanced AI products necessitate regulation, and company representatives present at the hearing conveyed a receptiveness to certain of the Council’s proposals.

"I share your commitment to ensuring that technology is developed safely and that its benefits are widely shared," said Morgan Dwyer of OpenAI.

At a certain juncture during the hearing, Julie Menin, the speaker of the City Council, enjoined each of the company representatives to “quantify the risk posed by AI in the worst case catastrophic scenario.”

Dwyer responded: “I do not know — nor, indeed, do I consider it material whether the probability of some catastrophic outcome stands at 1%, 10%, or 20%.

None of these magnitudes is even remotely tolerable.

We ought not to train models concerning which we cannot advance an exceptionally compelling case that we are capable of keeping them under human control.”

Menin excoriated the response as “flippant at best.”

"I would contend that the percentage chance is immaterial.

What is material is the commitment to safety, and that is precisely what OpenAI is committed to doing," Dwyer said.

The City Council is contemplating a suite of AI regulations, encompassing proposals that would mandate corporate disclosure regarding AI tools, stipulate third-party approval for AI systems deployed within the city, and entitle whistleblowers to a share of fines levied against AI companies found in violation of the law.

Concurrently, New York Governor Kathy Hochul, a Democrat, has undertaken to promulgate reporting and transparency standards governing AI, as have governors in other jurisdictions, albeit their authority remains somewhat circumscribed, given that AI technology transcends state boundaries and indeed spans the globe.

Logan Graham, representing Anthropic, apprised the Council that the company embraces “smart regulation” and that state and local governments have a part to play in the formulation of frameworks.

“AI is sufficiently potent that misjudging it carries tangible ramifications—for individuals, for institutions, for public trust,” he said.

“We must labour indefatigably to advance the science of risk mitigation.”