Narrowed AI Act scope risks fragmentation and leaves major consumer risks unaddressed

Following today’s adoption of the European Parliament’s position on the AI omnibus and ahead of interinstitutional negotiations, BEUC calls upon the EU institutions and national governments to ensure that critical safeguards to protect consumers and their fundamental rights are kept in the AI Act.

The Parliament’s proposal to delete a wide range of consumer-facing AI systems from the direct scope of the AI Act by no longer considering them as high-risk, could exclude many AI-embedded products – from children’s toys and smartwatches to medical devices – from essential compliance rules. This weakens consumer protections, as AI-specific risks of products covered by sectoral EU product legislation would go unaddressed. Instead of simplifying, this will create more fragmentation.

We welcome improvements to the original Commission’s AI omnibus proposal, such as keeping the AI Act's obligation to register high-risk AI systems. We also cautiously welcome the introduction of a new ban on generating sexually explicit deepfakes.

Agustín Reyna, Director General of BEUC, commented: 
“MEPs move to ban sexually explicit deepfakes addresses serious harms faced by consumers. But there is still much in the original proposal that must be improved. The AI omnibus as it stands still rolls back consumer protections and undermines fundamental rights, as it delays key safeguards, weakens transparency and creates major loopholes that narrow the scope of the AI Act. In doing so, the EU risks excluding many AI systems already embedded in consumer products, such as children's toys, smartwatches and medical devices, from the highest standards and putting consumers at even greater risk just months after the AI Act was agreed.”