OpenAI Safety Researcher Puts Human Extinction at 70% in 3 Years

Marcus Williams, a member of technical staff on OpenAI's Safety Oversight team, said on X that without AI regulation or a coordinated slowdown between frontier labs, human extinction in the next few years looks very likely. When pressed for a number, he put the odds at 70% within three years. He added that he considers regulation or a slowdown to be very possible.

What Williams actually said

A member of technical staff on OpenAI's Safety Oversight team put the odds of human extinction at 70% within three years unless AI labs slow down or regulators step in. Marcus Williams, who joined the company in May 2025, made the call on X. He posted the warning first, then attached a number when another user pushed him for one.

Another user then turned it into a wager, offering $10,000 in Microsoft shares against OpenAI stock over five years. Williams answered with a percentage. The exchange has since been amplified across AI safety and prediction-market circles.

Context: a week of safety blowback

The post lands in a week that pushed lab safety into mainstream politics. On September 9, an Anthropic researcher resigned and accused Anthropic and OpenAI of gambling with human lives. US lawmakers responded within a day. A fresh push to pause AI followed, including a Senate bill that would ban superintelligence outright.

Voices outside the labs have echoed the alarm. The UN human rights chief called advanced AI an existential risk earlier this week in Geneva.

Do you want to see how to make more plays? Do you want to find gains yourself?

Unusual Whales helps you find market opportunities through our market tide, historical options flow, GEX, and much, much more.

Create a free account here to start conquering the market with Unusual Whales.

Where Williams sits vs. peers

Williams is on the aggressive end of extinction forecasts from inside the labs. Evan Hubinger, who leads alignment science at Anthropic, puts the odds above 10% this decade. Nobel laureate Geoffrey Hinton has cited a range of 10% to 20%.

Meanwhile, crypto traders now track these warnings directly. Polymarket pushed Williams's post to its own followers within hours, and its bettors already price elections and macro events alongside token odds.

Why it matters for the tape

Safety-driven regulation is the single biggest non-macro overhang on the AI trade. A Senate proposal to ban superintelligence, combined with high-profile researcher defections, raises the odds of compute caps, licensing regimes, or export-style controls on frontier training runs.

Any of those would hit model-layer capex and the picks-and-shovels names that have carried the index. The market has, so far, ignored the safety chatter. That gap is the trade.

Options market and stocks to watch

MSFT: OpenAI's largest backer and the direct counterparty in the $10,000 wager tied to Williams's post. Watch for headline sensitivity around any OpenAI governance or safety escalation.

NVDA: Any regulatory push toward training-compute caps or licensing lands here first. Watch flow around policy headlines.

GOOGL: A frontier-lab operator via DeepMind. Watch for spillover if a coordinated slowdown narrative gains traction.

META: Open-weights strategy makes it a target in any superintelligence-ban framework. Watch for policy-driven repricing.

AMD: Second-order beneficiary or victim of any compute-cap regime alongside NVDA.

For more, see other news on AI policy and safety.

Want more market intelligence? Create your free Unusual Whales account for options flow, market tide, GEX, and the full toolkit.