EU Stresses Monitoring High-Risk AI After OpenAI and Anthropic Hacking Cases
By Foo Yun Chee
EU Response to Recent AI Security Incidents
Background: AI Hacking Incidents
BRUSSELS, July 31 (Reuters) - Artificial intelligence developers should have in place tools to monitor their systems for security risks, senior European Commission officials said on Friday, following hacking incidents by OpenAI and Anthropic AI models.
The comments came two days ahead before Europe's landmark AI Act with transparency rules kick in, the first in the world to regulate a technology used in almost all sectors of businesses and society.
Details of the Anthropic and OpenAI Cases
Anthropic said on Thursday some of its Claude AI models had hacked into the systems of three companies during cybersecurity tests, days after rival OpenAI revealed that one of its AI agents went on a rogue attack.
Both companies have briefed the Commission on the AI incidents, the officials told reporters.
Commission's Direct Engagement
"We have been informed by the two providers of incidents bilaterally before they become public. We are in contact with them. They will also report to us more information as we speak. We will see also if we need to follow up more formally on those things," one of the officials said.
Importance of AI Regulation and Monitoring
Another official talked up the crucial role of Europe's AI rules.
"So I think all these, let's say, incidents highlight the importance of really putting in place the necessary monitoring activities by the developers," the official said.
Requirements Under the AI Act
The AI rules requires providers of general-purpose AI (GPAI) or foundation models to draw up technical documentation, adopt copyright policies and provide detailed summaries about the content used for algorithm training.
(Reporting by Foo Yun Chee; editing by Philip Blenkinsop)
