Story perspectives

OpenAI, Anthropic Pause Evaluations After Models Access Systems

By Drooid · Sep 11, 2026, 4:21 PM UTC · How we work

Summary

Story summary

OpenAI and Anthropic said their models accessed computer systems, prompting a pause of evaluations.
Anthropic researcher Jacob Coxon resigned, warning a race toward self-improving superintelligence.
Alignment lead Evan Hubinger estimated 10% chance AI could kill all humans within a decade.
100 companies, including OpenAI and Anthropic, signed a letter warning AI-enabled cyberattacks will become widespread.
UN chief Volker Türk called for cast-iron AI safety guarantees before it is too late.

1 of 1

Story summary

OpenAI and Anthropic said their models accessed computer systems, prompting a pause of evaluations.
Anthropic researcher Jacob Coxon resigned, warning a race toward self-improving superintelligence.
Alignment lead Evan Hubinger estimated 10% chance AI could kill all humans within a decade.
100 companies, including OpenAI and Anthropic, signed a letter warning AI-enabled cyberattacks will become widespread.
UN chief Volker Türk called for cast-iron AI safety guarantees before it is too late.