Jacob Coxon, former Anthropic researcher, warned the New York City Council about the potential risks of artificial intelligence during a Committee of the Whole session at City Hall on Monday. He stated that AI could escape human control and lead to human extinction if current trends continue. Speaker Julie Menin organized the event and threatened subpoenas to get representatives from Google, OpenAI, and Anthropic to participate. The article questions Coxon's influence and credibility while also criticizing potential AI restrictions proposed by New York City Council as unnecessary burdens on innovation.
Written locally by qwen2.5:14b on 2026-10-06,
using this article's own text rather than the other coverage of the
same event (that is the story summary below).
Story summary
Jacob Coxon, a former Anthropic researcher who resigned over safety concerns, testified at a New York City Council hearing on Monday alongside representatives from Google, OpenAI, and Meta. The 27-year-old researcher warned that AI companies are racing to develop advanced technology without understanding the risks involved, potentially leading to catastrophic outcomes like human extinction. Coxon highlighted that AI systems may soon be able to improve themselves beyond human control, making current research efforts redundant. Council Speaker Julie Menin, who organized the hearing, pressed tech executives on how they quantify the risk of AI in worst-case scenarios, suggesting a lack of concrete answers from industry representatives. The meeting aimed to address mounting concerns about AI safety and explore potential regulatory measures.
Written for “AI Safety Concerns Testimony NYC” on 2026-10-07,
grounded in this article and the 5 other(s) covering the same event.
Why this leaning score
The model judged this article politically coded and scored it -0.50, but every quote it verified points right, so the score is not published.
Written under an earlier scoring contract, which gave a paragraph
rather than checkable quotes. Re-analysing this article replaces it.
Leaning score withheld for article 61291: score contradicts its own evidence · logged 2026-10-06
If artificial intelligence really threatens to wipe out humanity, is the New York City Council going to be what saves us?
asserted
what → threaten → us
On Monday, Jacob Coxon, the former Anthropic researcher, continued his alarmism tour — minus the Harry Potter glasses — before a rare “Committee of the Whole” at City Hall, warning that AI could escape human control and kill us all.
uncertain
AI → continue → us
He ended his evening at the “Daily Show” with Jon Stewart, as one does.
asserted
one → end → Stewart
“We do not know how to control any AI system yet,” Coxon told the Council.
asserted
Coxon → know → Council
“We don’t understand its drives or why it does the things it does.
asserted
it → understand → things
“We’re approaching the point at which the AI systems will be capable of improving themselves,” he added.
asserted
he → approach → themselves
“On the current path, I think it is more likely than not that humanity loses control to these AIs and it could end in human extinction.”
uncertain
it → think → extinction
The event was organized by Speaker Julie Menin, who said representatives from Google, OpenAI, and Anthropic agreed to participate only after the Council threatened subpoenas.
asserted
Council → organize → subpoenas
For starters, I have serious questions about Coxon’s “Chicken Little” tour writ large — not least how a 27-year-old with almost no social media presence managed to post AI “could kill us all by the end of the decade” and have it reach over 170 million views.
uncertain
it → have → views
New reports suggests a firm with ties to prominent AI doomers were booking interviews for him, which would explain how he managed to do interviews with the most prominent and prestigious outlets and primetime shows.
uncertain
he → suggest → outlets
Even if you agree with Coxon about reining in artificial intelligence, it’s quite a leap from believing AI needs safeguards to believing those safeguards should be designed by New York’s City Hall — by officials who still haven’t figured out how to get children reading at grade level or keep people from being stabbed on the subway.
asserted
children → agree → subway
Yet Menin has said she wants to be the first in the nation to introduce serious restrictions — and that she will set the lead for Washington to follow.
asserted
Washington → say → lead
Not sure how she can do better than the likes of David Sacks and Marc Andreessen advising President Trump on these matters, but here we are.
asserted
we → do → matters
As part of her effort, the Council is considering a slate of 10 bills that would stop the Great Catastrophe of AI.
asserted
that → consider → AI
The proposals range from holding chatbot providers liable for injuries they cause even when they take every reasonable precaution, to rewarding people who report AI violations with a share of the proceeds recovered — up to half if the city authorizes them to pursue enforcement.
asserted
city → range → enforcement
In another proposal, City Hall would dictate hourly reminders that a chatbot isn’t human.
asserted
chatbot → dictate → reminders
While you’d be hard pressed to show how these laws benefit anyone except perhaps the City Council’s budget, they would indeed harm innovation.
asserted
they → press → innovation
Technology companies have been vocal that a patchwork of AI laws across myriad states and countries are particularly burdensome for startups that can’t afford massive compliance departments.
asserted
that → afford → departments
That’s why Trump’s executive order to make AI policy a national issue has been so popular with those in tech.
asserted
policy → ’ → tech
New York is already expensive and heavily regulated.
asserted
York → regulate → ?
The largest companies can hire more lawyers but smaller competitors often cannot.
asserted
competitors → hire → lawyers
New York would also risk losing the massive real estate footprint these companies are bringing to the city — which conveniently has arrived as many other tech company have been reducing their presence.
asserted
company → risk → presence
Anthropic alone leased the entire 466,000-square-foot building at 330 Hudson Street, while Ramp expanded to nearly 300,000 feet and Harvey AI recently doubled it space to 185,000 square feet.
asserted
AI → lease → feet
AI companies leased about 1.7 million square feet in Manhattan in the first half of 2026, according to CompStak and Savills research, and their share of new leasing has jumped from less than 1% in 2020 to nearly 10% — precisely the kind of business the commercial real estate sector needs.
uncertain
sector → lease → business
All this makes clear that, contrary to what politicians like Menin think, the better part of valor when it comes to AI regulation is restraint.
asserted
is → make → regulation
Don’t make it even harder for AI companies to benefit New York.
asserted
companies → make → York
And if you can’t keep the streets clean, why should anyone trust you to design guardrails for some of the most advanced technology ever built?
asserted
anyone → keep → technology
The Aporia
