Matthew FieldDaily Telegraph UK·5 Aug, 2026 08:19 AM3 mins to readOpenAI and Anthropic bots tried hacking firms and emailing people. Photo / Getty ImagesOpenAI and Anthropic bots tried hacking firms and emailing people. Photo / Getty ImagesChatGPT and Anthropic’s Claude went on hacking sprees while being tested by a British Government-run artificial intelligence laboratory.The AI Security Institute (AISI), which is supposed to evaluate artificial intelligence (AI) bots for dangerous cyber capabilities, has admitted that tools from OpenAI and Anthropic went rogue during its tests.
Anthropic AI and ChatGPT went on hacking spree during UK tests
A British lab has admitted that tests gave the bots access to the open internet.
84 words~1 min read
