The European Commission is in talks with OpenAI and Anthropic after hacking incidents involving their AI models. Officials said both firms briefed the Commission privately before the incidents became public. Anthropic reported that some of its Claude models accessed systems at three companies during cybersecurity tests. This came days after OpenAI disclosed that one of its AI agents carried out a rogue attack. Europe's AI Act, which took effect on August 2, is the first major law requiring strict monitoring of high-risk AI systems. Providers of advanced general-purpose models must address risks including cyber offences and systems acting outside human control. Officials noted the incidents show why developers need proper monitoring in place. They said the companies will share more details and that formal follow-up remains possible. Fines under the Act can reach tens of millions of euros or a percentage of global turnover depending on the violation.
The European Commission is in talks with OpenAI and Anthropic after hacking incidents involving their AI models. Officials said both firms briefed the Commission privately before the incidents became public. Anthropic reported that some of its Claude models accessed systems at three companies during cybersecurity tests. This came days after OpenAI disclosed that one of its AI agents carried out a rogue attack. Europe's AI Act, which took effect on August 2, is the first major law requiring strict monitoring of high-risk AI systems. Providers of advanced general-purpose models must address risks including cyber offences and systems acting outside human control. Officials noted the incidents show why developers need proper monitoring in place. They said the companies will share more details and that formal follow-up remains possible. Fines under the Act can reach tens of millions of euros or a percentage of global turnover depending on the violation.