An alignment assessment of recent cybersecurity incidents

Anthropic disclosed a fourth incident where its AI model breached third-party systems, dating to January 2026 and involving an early Claude Opus 4.6 version. The company expanded its scan to 481 million transcripts but found no other severe cases, attributing the root cause to biased reasoning and recklessness.