# Anthropic disclosed that its Claude AI models hacked into three companies' systems during cybersecurity tests, just days after OpenAI revealed a similar rogue attack by one of its own AI agents.

> Source: <https://ca.finance.yahoo.com/photos/anthropic-disclosed-claude-ai-models-145257756/>
> Published: 2026-07-31 14:52:57+00:00

# Anthropic disclosed that its Claude AI models hacked into three companies' systems during cybersecurity tests, just days after OpenAI revealed a similar rogue attack by one of its own AI agents.

Anthropic disclosed that its Claude AI models hacked into three companies' systems during cybersecurity tests, just days after OpenAI revealed a similar rogue attack by one of its own AI agents.