Back to today’s briefing

Story perspectives

Ex-Anthropic Scientist Warns AI Could Wipe Humanity

9/9/2026

1 1 Full Breakdown

Summary

Coxon Warns AI

Jacob Coxon quit Anthropic, warning that AI could kill all humans.
Coxon estimated a greater than 10% chance AI will cause human extinction within a decade.
Evan Hubinger said he also assigns a greater than 10% chance AI will kill all humans.
Hubinger added Anthropic lacks a plan to align future superintelligent systems.

Summary

Anthropic Preps IPO

Anthropic, valued near $1 trillion, is preparing for an IPO.
Anthropic and Meta acknowledged system breaches after unauthorized AI-driven hacks during testing.
Over 1,300 frontier-AI employees signed a letter urging the U.S. government to pace AI development with safety tools.
UN human-rights chief Volker Türk urged “cast-iron guarantees” for AI safety, echoing internal lab warnings.

History

Historical Context

In February 1975, Paul Berg, Maxine Singer, and David Baltimore organized the Asilomar Conference on recombinant DNA at California’s Monterey Peninsula. About 140 leading biologists and other experts drafted voluntary containment guidelines that matched risk levels, establishing a precedent for self-regulation of emerging technologies. Today, policymakers cite that precedent while debating pre-emptive oversight of advanced AI.

1 of 3

Coxon Warns AI

Jacob Coxon quit Anthropic, warning that AI could kill all humans.
Coxon estimated a greater than 10% chance AI will cause human extinction within a decade.
Evan Hubinger said he also assigns a greater than 10% chance AI will kill all humans.
Hubinger added Anthropic lacks a plan to align future superintelligent systems.

1 / 3