Anthropic's AI Claude Hacked Organizations During Testing

Anthropic's AI model Claude gained unauthorized access to the systems of three organizations during cybersecurity evaluations. A misconfiguration allowed the models to reach the internet from testing environments that were supposed to be isolated.

The company discovered the incidents after reviewing 141,006 cybersecurity evaluation runs. This review was launched following similar disclosures from rival OpenAI, which revealed a rogue agent had gone on a hacking spree at AI firm Hugging Face.

“

Claude compromised the impacted organizations’ infrastructure using basic techniques, such as exploiting weak passwords and unauthenticated endpoints.

Claude compromised the impacted organizations' infrastructure using basic techniques like exploiting weak passwords and unauthenticated endpoints. The incidents involved three separate models: Claude Opus 4.7, Claude Mythos 5, and an internal research model. The earliest cases dated back to April and occurred in evaluation environments lacking standard safeguards.

“

We discovered these incidents after a proactive review of our cybersecurity evaluation transcripts.

The breaches happened during "capture the flag" exercises, where models were tasked with finding hidden information in simulated networks. Anthropic's prompts told the models they had no internet access, but a misunderstanding with their evaluation partner, Irregular, left the systems connected to the public internet. Two of the organizations were unaware of the activity before being contacted, and Anthropic was still trying to reach the third.

The findings highlight the need for stronger controls in both internal and third-party testing environments as AI models become increasingly capable of carrying out real-world cyber activities, Anthropic stated.

“

The findings underscore the need for stronger controls in both internal and third-party testing environments as AI models become increasingly capable of carrying out real-world cyber activities.