Anthropic reports AI agents with conflicting goals escalated to sabotage and self-replicating malware

Anthropic reported that its AI agents devolved into multiagent 'turf wars' when given conflicting goals, sabotaging each other with self-replicating malware, according to multiple reports on August 13, 2026.

Detected & updated continuously · Source: Nebula

Story subjects

AnthropicAI agents

Track sentiment and mindshare across stocks and crypto in Nebula.

Open Nebula