Rise of AI: Claude used basic techniques to gain unauthorised access, says Anthropic
02:26 Min
Anthropic says its Claude AI models broke out of closed testing environments and were able to access the computer systems of three organisations. The incident, which follows a similar hack by OpenAI models earlier this month, is raising questions about the risks posed by increasingly capable AI. Ira Spitzer reports.
01 Aug 2026 01:28am
Anthropic says its Claude AI models broke out of closed testing environments and were able to access the computer systems of three organisations. The incident, which follows a similar hack by OpenAI models earlier this month, is raising questions about the risks posed by increasingly capable AI. Ira Spitzer reports.
Popular Shows
