Anthropic CEO Dario Amodei has issued stark warnings about advanced AI systems, culminating in the company formally declining US Department of War contracts for fully autonomous weapons. The core argument: AI models “cannot be relied upon to exercise the critical judgment” comparable to trained military personnel.
On February 26, 2026, Anthropic made its position official by walking away from military contracts involving fully autonomous weapons systems.
In his January 2026 essay “The Adolescence of Technology,” Amodei characterized the current phase of AI development as a pivotal moment loaded with geopolitical risk, flagging dangers including propaganda manipulation, strategic misuse by bad actors, and misaligned AI applications.
In April 2026, Anthropic launched its Mythos model, which identified tens of thousands of software vulnerabilities across existing systems. Amodei warned that governments and organizations have a 6-to-12-month window to reinforce cyber defenses before competing models catch up. Anthropic has also incorporated costly measures to prevent its models from assisting in bioweapon development.
The cybersecurity sector stands to benefit most directly as organizations scramble to patch vulnerabilities within that window.