Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsAt an October 5, 2026, New York City Council hearing, former AI-lab employees warned that increasingly capable systems could outpace human control. Representatives of OpenAI, Anthropic, Google and Meta described safety efforts and responded to proposals for stronger oversight. The hearing captured a sharp disagreement; it did not establish that catastrophe is inevitable or produce a measured probability of AI-related catastrophe.
What happened at the October 5 hearing?
The New York City Council convened a Committee of the Whole hearing, bringing together all 51 council members to examine risks from rapidly developing AI and possible safeguards. The Council’s September 16 announcement described the hearing’s purpose and the city’s existing AI oversight framework.
As an Amazon Associate I earn from qualifying purchases.
Associated Press coverage of the hearing reported testimony from representatives of OpenAI, Anthropic, Google and Meta, alongside former employees of AI companies. The Council’s earlier September 28 notice had described a different stage of the process: it said the companies other than Meta confirmed after the Council threatened subpoenas, and that SpaceXAI had not responded and a subpoena had been issued. That advance account should not be confused with AP’s later report of who participated in the hearing.
What did former AI-lab employees warn?
Jacob Coxon, whom AP identified as a former Anthropic engineer, made the hearing’s starkest prediction: “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.” That is Coxon’s view, as reported by AP—not a finding by the Council or an established risk estimate.
#1 Best Overall
Coxon argued that the pace and culture of development were part of the problem. “The companies run on a startup mindset: Move fast, break things, fix them later. That works for a photo sharing app. It does not work for building the most powerful technology ever built,” he said, according to the Associated Press account.
Anthropic representative Logan Graham also acknowledged the stakes, saying: “AI is powerful enough that getting this wrong has real consequences, for individuals, institutions, public trust. We have to work relentlessly to do the science of mitigating the risks.”
Rank #2
Did OpenAI give a probability for catastrophic AI risk?
No. When Council Speaker Julie Menin asked representatives to quantify catastrophic risk, OpenAI representative Morgan Dwyer did not offer a single estimate. He replied: “I don’t know – I also don’t think it matters whether it’s 1% or 10% or 20% chance that something catastrophic will go wrong. None of these levels is remotely acceptable. We should not train models that we cannot make an extremely strong case that we can keep under human control.”
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
The percentages were hypothetical examples in Dwyer’s answer, not measured estimates or an OpenAI forecast. AP reported that Menin called his response “flippant at best”; Dwyer then emphasized the company’s commitment to safety. He also told the council: “I share your commitment to ensuring that technology is developed safely and that its benefits are widely shared.”
How did the positions differ?
The former employees’ warning was that development may be moving faster than the ability to control advanced systems. Company representatives emphasized safety work and the need to mitigate risks; Dwyer argued that even a low probability of catastrophe would be unacceptable. The exchange did not settle how likely loss of control is, what evidence would demonstrate adequate control, or whether a particular risk percentage can be meaningfully established.
The disagreement therefore concerned both the assessment of danger and what should follow from uncertainty: Coxon called for taking the loss-of-control possibility seriously, while Dwyer framed the standard as refusing to train models without a strong case for keeping them under human control. The hearing report does not establish a shared probability or a consensus on what evidence meets that standard.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What AI oversight does New York City already have?
Before the hearing, New York City had enacted AI-related oversight measures in 2025. The Council says those laws created an Office of Algorithmic Accountability to review AI and algorithmic systems used by city agencies, conduct risk assessments, investigate potential harms and require corrective action when appropriate.
The Council also describes standards for agency AI development, procurement and use that cover privacy, civil rights, security, monitoring and risk management, along with a public inventory of covered systems. These measures concern the city’s own agencies; they are distinct from proposals discussed at the hearing for broader safeguards.
Best Value
What safeguards were proposed?
The Council’s September 28 notice and AP’s hearing report describe proposals under consideration, not measures verified as enacted after the hearing. They included:
- A whistleblower incentive or reward program.
- A private right of action for New Yorkers harmed by AI agents.
- Independent third-party validation requirements.
- Disclosure requirements for AI tools, as reported by AP.
The available accounts do not establish whether these proposals were later amended, passed or signed into law. The hearing also did not resolve the limits of what city government can require of AI companies operating beyond municipal services.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
