EU Mandates Pre-Market Checks for AI Models from OpenAI, Google, and Others

The EU is moving to evaluate Big Tech AI models before their release. Starting Tuesday (local time), the European Union is exercising strong oversight powers to evaluate and restrict the performance and safety of AI models from Big Tech companies like OpenAI and Google prior to their market launch. This enforcement is part of the phased implementation of the EU AI Act established in 2024. Executive Vice-President Henna Virkkunen stated that the move is intended to control the risks posed by state-of-the-art AI models.

The EU is regulating all companies offering AI models in Europe, regardless of their headquarters' location, mandating the designation of an official representative in Europe and the submission of information. In case of regulation violations or refusal to submit information, a fine of up to 3% of annual turnover or 15 million euros, whichever is higher, will be imposed on the company.

Consultations are already underway, with OpenAI and Anthropic having contacted the EU AI Office. OpenAI and Google have expressed their commitment to comply with EU regulations and continue cooperation, responding to the new regulatory environment.