Legal experts for the European Commission and the European Parliament are locked in a dispute over the interpretation of key provisions of the AI Act, especially how these rules are reflected in the code of practice for AI models. The conflict centers on how providers of powerful AI models — such as OpenAI and Google — should assess and manage societal risks. Parliamentary officials argue that the code underestimates fundamental rights and democracy concerns by narrowly defining systemic risks and prioritizing long-term catastrophic scenarios. Tensions escalated after informal bilateral discussions took place with the independent experts chairing the code’s drafting process, leading to a call for joint meetings between the two legal teams.