Story perspectives

AI Tsunami Warning: Anthropic Abandons Safety Pledge Amid Pressure

3/1/2026

Summary

Story summary

Dario Amodei, Anthropic Chief Executive Officer, warned that an AI tsunami could disrupt society as technology approaches human-level intelligence.
He criticized the lack of recognition of this threat.
Anthropic abandoned its safety pledge, allowing AI model training without guaranteed guardrails.
The shift occurred under Pentagon pressure to relax restrictions linked to a $200 million defense contract.
Critics say the move undermines Anthropic's safety-focused image and raises concerns about AI risk.

History

Historical Context

The U.S. established the Defense Innovation Unit in 2016 to enhance AI development for national security, following DARPA's historical role in advancing technology since its inception in 1958. Dario Amodei's warnings about AI risks reflect ongoing concerns about balancing innovation with ethical governance, echoing past debates over technological advancements outpacing regulatory frameworks, particularly after the September 11 attacks and the creation of controversial surveillance programs.

1 of 2

Story summary

Dario Amodei, Anthropic Chief Executive Officer, warned that an AI tsunami could disrupt society as technology approaches human-level intelligence.
He criticized the lack of recognition of this threat.
Anthropic abandoned its safety pledge, allowing AI model training without guaranteed guardrails.
The shift occurred under Pentagon pressure to relax restrictions linked to a $200 million defense contract.
Critics say the move undermines Anthropic's safety-focused image and raises concerns about AI risk.

1 / 2