Europe issues draft high-risk AI rules

The European Commission moved on May 19 to answer one of the most contested questions in the European Union’s AI Act: which systems count as “high-risk.” The draft guidance, published by the Commission’s digital policy arm, is meant to help providers, deployers and market surveillance authorities decide when stricter obligations apply under Article 6 of the law. Brussels also opened a targeted consultation that runs through June 23, asking for feedback on both the text and the practical examples that accompany it. The timing matters because companies are trying to build products and compliance programs while key parts of the EU rulebook are still being translated into operational tests. The Commission said the draft guidelines are intended to support “uniform application and effective enforcement” of Article 6. IAPP, which reviewed the documents, said the package is organized as a three-part guide covering general principles and the two high-risk categories set out in the AI Act. (digital-strategy.ec.europa.eu) ### Which part of the AI Act is Brussels trying to clarify? Article 6 is the hinge. The Commission’s draft says the guidelines are designed to help determine whether an AI system should be classified as high-risk, which in turn decides whether providers and deployers face the AI Act’s stricter obligations. Those obligations attach to systems in the categories listed by the law, rather than to every AI product sold in Europe. (digital-strategy.ec.europa.eu) The Commission split the guidance into three documents. One covers general principles, a second addresses systems linked to Annex I of the AI Act, and a third addresses Annex III use cases. The consultation notice says the examples are meant to show how classification should be assessed “in different areas and use cases.” (digital-strategy.ec.europa.eu) ### Why did the Commission publish examples alongside the draft? Practical examples are part of the answer because the legal text leaves room for argument at the edges. The Commission said it issued “a list of practical examples for stakeholder feedback” together with the draft guidelines. IAPP reported that the examples are intended to illustrate how the classification should be assessed across sectors and use cases. (digital-strategy.ec.europa.eu) June 23 is the next formal checkpoint. The consultation page says feedback will be collected from May 19 through June 23, after which the Commission can revise the text before finalizing guidance. ### How does Meta’s WhatsApp offer fit into the same story? (digital-strategy.ec.europa.eu) Meta Platforms was dealing with a different European pressure point at the same time: access. Reuters reported on May 19 that Meta offered rival AI chatbots, including OpenAI, free access to WhatsApp in Europe up to a limit, after which fees would apply. Two people with knowledge of the matter told Reuters the proposal was made as Meta sought to address concerns from EU antitrust regulators. (digital-strategy.ec.europa.eu) WhatsApp matters because distribution is becoming part of the AI fight. Reuters reported that Meta’s offer centered on Europe and would apply to rival chatbot services using the messaging platform, with charging beginning once usage crossed a threshold. ### Why is OpenAI still part of the Europe conversation after winning in court? (msn.com) OpenAI won its case against Elon Musk in Oakland, California, but the trial still exposed internal disputes that could affect how regulators and partners view the company. AP reported that Musk lost his challenge against OpenAI and its top executives after seeking changes that included Sam Altman’s ouster. The same report said witness testimony included accusations that Altman had lied or been dishonest in past dealings. (msn.com) Those courtroom details landed as OpenAI remains a central company in European AI debates over competition, access and compliance. AP said the verdict spared OpenAI a costly legal setback, while Reuters reported separately that testimony from former colleagues could leave lasting reputational damage for Altman despite the courtroom win. (apnews.com) ### What happens next in Brussels? June 23 is the deadline for comments on the Commission’s draft high-risk guidance. The feedback process is being run through the Commission’s digital policy portal, where the three draft documents and consultation notice were published on May 19. Meta, OpenAI and other AI companies now have a live window to argue over how Europe defines risk while separate disputes over platform access and competition continue. (apnews.com) (digital-strategy.ec.europa.eu)