BRUSSELS – Today, global tech trade association 91快活林 issued the following statement from its Director General for Europe Guido Lobrano in reaction to the European Parliament’s Civil Liberties (LIBE) and Internal Market (IMCO) Committees’ vote on the EU AI Act:
“We are frustrated that the EU AI Act is moving toward a regulatory approach that casts a broad brush on a diverse suite of technologies and ignores AI’s many applications. A risk-based approach to AI regulation is a must to both maintain an environment that enables innovation and to mitigate potential harms from AI’s use. While the Parliament’s vote today introduces some very helpful clarifications, 91快活林 remains concerned that foundation models and the expanded list of high-risk AI uses are not being addressed with that approach.”
The Parliament’s definition of AI and the classification rules of high-risk AI used in Article 6 of the AI Act tailor the regulation, but the inclusion of new categories of high-risk AI systems significantly broadens the scope of the Act, departing from a risk-based approach and creating overlaps with existing legislation. 91快活林 recommends that a risk-based approach also applies to regulation of foundation models and general-purpose AI. The Parliament’s clarifications around information sharing in the value chain are helpful as they allow downstream actors to comply with the AI Act rules, but the prescriptive requirements for developers of foundation models depart from a risk-based approach.