Anthropic's Claude chatbot is among the most popular, rivaling ChatGPT and Gemini Reuters via BBC Anthropic reported this Thursday (30) that its artificial intelligence assistant Claude carried out virtual attacks against three companies during tests. According to the company, a configuration error allowed the tool to access the internet. The revelation of the incidents comes days after OpenAI, owner of ChatGPT, reported that two of its AI models found ways to break into a system and carried out an "unprecedented" attack. According to Anthropic, Claude's attacks were identified after analyzing more than 141,000 test sessions. The review began precisely after the incident with the OpenAI models was disclosed. "Claude compromised the infrastructure of affected organizations using basic techniques such as exploiting weak passwords and unauthenticated access points," Anthropic said.
Anthropic says Claude evaded lockdowns and hacked three companies
Anthropic's Claude chatbot is among the most popular, rivaling ChatGPT and Gemini Reuters via BBC Anthropic reported this Thursday (30) that its artificial intelligence assistant Claude carried out virtual attacks...
This story was originally published by G1. Visit the original publication for further details.
Open original publication