The Intelligence from The Economist
The Intelligence from The Economist

The end of the world is AI? An existential threat

September 16, 2026

AI Summary

5 min read

A 10% chance of everything ending? Inside the AI extinction debate

In early September 2025, Jacob Coxson, a former researcher at both OpenAI and Anthropic, resigned loudly and publicly with a warning: AI safety is an imminent and potentially existential threat to the world. His former colleagues largely agreed. Evan Hubinger, who leads Anthropic's AI safety team, put his own estimate at more than 10% — an "astronomically high figure when you're talking about the destruction of everything," as The Economist's AI writer Alex Hearn put it. That number has pushed the question of human extinction from fringe futurist circles into the center of the AI conversation.

Why the alarm is growing now

The fear of runaway AI is not new. For years, a thread in the AI world has looked at continued rapid improvements in systems and extrapolated what happens if that improvement doesn't stop. What changed in 2025 is that for the last five-plus years, the improvement hasn't stopped, and there is "less and less runway" before you either work out what stops AI getting better or end up with AI that is "very unsettlingly competent."

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

What you'll learn

  • 1 (01:34) **The Existential AI Threat Debate** - Hosts introduce the central question: is AI an extinction risk for humanity?
  • 2 (03:37) **Alex Hern Introduces the Core Argument** - The Economist's AI writer explains why the extinction risk conversation has escalated in 2026.
  • 3 (04:29) **Why the Alarm Has Grown in 2026** - The conversation shifts to explain the long-running thread of concern that has now become acute.
  • 4 (05:19) **The Hugging Face Attack: A Concrete Trigger** - Alex Hern details the specific incident that shocked the AI world and made the threat tangible.
  • 5 (06:50) **Can AI Development Be Slowed?** - The discussion moves to the feasibility of a pause, framing it as a game-theoretic and governmental problem, not a technical one.
  • 6 (08:48) **The Geopolitical Race: US vs. China** - The conversation examines how the US-China rivalry intensifies the pressure to develop AI without restraint.
  • 7 (10:15) **What a Slowdown Would Actually Look Like** - The host asks for a concrete picture of what a pause or slowdown entails beyond abstract calls.

+ Full timestamped outline available in the app

Guests on this episode

Evan Hubinger

Show Notes

Is artificial intelligence becoming uncontrollably powerful? Our correspondent analyses whether a pause to breakneck development is possible or feasible. We have entered the new age of the orator. And why cheese is getting more popular in India. 


Guests and host:

  • Alex Hern, AI writer
  • Tom Wainwright, culture correspondent
  • Tom Sasse, South Asia bureau chief
  • Rosie Blau, co-host of “The Intelligence”
  • Jason Palmer, co-host of “The Intelligence”


Topics covered: 

  • Human extinction, AI, Sam Altman, Dario Amodei
  • Oral culture, literacy, James Marriott, Walter Ong
  • Cheese, India, kalimpong


Listen to what matters most, from global politics and business to science and technology—subscribe to The Economist.



Hosted on Acast. See acast.com/privacy for more information.

The Intelligence from The Economist

More from this podcast

The Intelligence from The Economist →