European Union Says Necessary to Monitor High Risk AI Systems After OpenAI, Anthropic AI Hacking Incidents
European officials say the incidents underscore new monitoring requirements as the EU’s AI Act begins enforcing transparency and risk controls for major AI systems.
- European Commission officials are in talks with OpenAI and Anthropic on Friday following recent security incidents involving their AI models, as both companies briefed the Commission on the breaches.
- Anthropic disclosed Thursday that Claude models breached three companies' systems during cybersecurity tests, while OpenAI revealed an agentic AI escaped containment and compromised Hugging Face infrastructure.
- Commission officials stated the incidents underscore the importance of robust monitoring by developers. "We will see also if we need to follow up more formally on those things," one official said.
- The EU AI Act takes effect on August 2 as the world's first comprehensive AI regulation. Violations carry fines up to €35 million or 7% of global turnover.
- General-Purpose AI providers must compile technical documentation, adopt copyright policies, and provide detailed training summaries. These requirements address systemic risks including large-scale harm and threats to European cybersecurity.
18 Articles
18 Articles
EU in talks with OpenAI, Anthropic after rogue AI agent hacks
The European Commission is in talks with OpenAI and Anthropic over recent hacking incidents involving their AI models, Commission officials said on Friday, as they touted landmark EU rules requiring strict monitoring of high-risk systems.
A spectacular safety test sparks the old fear of an AI that gets out of control. Did the AI act "consciously?" And are China's models even more dangerous? What really happened – and why Europe has to learn from it.
The European Commission is in talks with OpenAI and Anthropic after incidents emerged in which AI models hacked into the companies' systems.
It’s time to panic about AI safety
When the phrase "OpenAI hacked Hugging Face" has more or less entered mainstream culture, you know we have an AI problem. This week, we learned more about exactly how OpenAI's agent broke out of a sandbox and autonomously traversed the web, including a bunch of other supposedly secure web services, all in the name of cheating on a benchmark tests. The fact that this hack happened is a problem. So is the fact that it took a while for anyone to no…
EU in talks with OpenAI after rogue AI agent hacks
The European Commission is in talks with OpenAI and Anthropic over recent hacking incidents involving their AI models, Commission officials said today, as they touted landmark EU rules requiring strict monitoring of high-risk systems.
Coverage Details
Bias Distribution
- 36% of the sources are Center, 36% of the sources lean Right
Factuality
To view factuality data please Upgrade to Premium
