The European Commission has published new guidelines for companies operating AI models deemed to pose systemic risks.

The guidelines are intended to alleviate concerns of AI developers and businesses about the regulatory requirements of the AI Act, while providing greater clarity on how to comply with the law's requirements. Under them, companies that fail to comply could face significant penalties, ranging from €7.5m or 1.5% of annual turnover up to €35m or 7% of global turnover. The AI law, passed last year, will start to apply on August 2 for AI models deemed to have systemic risks, including basic general-purpose models developed by companies such as Google, OpenAI, Meta, Anthropic and Mistral.