The European Commission is expected to propose new transparency requirements for developers of large foundation models next week, according to two people familiar with the plan. The draft would mark another step in Brussels’ push to bring major AI systems under closer scrutiny.

Under the proposed rules, companies building advanced models would likely have to share summaries of the data used to train them and provide estimates of their energy consumption. Supporters of the approach argue that greater disclosure can help regulators, researchers, and the public better understand how these systems are built and what their environmental cost may be.

The move comes as governments around the world race to set guardrails for fast-moving AI technology. European policymakers have aimed to position the bloc as a global rule-maker on digital policy, while also responding to concerns about accountability, safety, and market power in the hands of a few large developers.

The draft is not yet final, and details could still change before publication. If adopted, the rules would add to the growing body of EU oversight covering artificial intelligence, data use, and corporate transparency.