AI is already in your child’s toy, your doctor’s device, and on your wrist. Now Brussels is deciding whether it needs the same strict rules as other high-risk systems, or whether existing product laws are enough. Trilogue talks on the AI Digital Omnibus conclude on Tuesday.
Industry has complained about a ‘double regulation’. Companies already complying with EU product laws would also have to meet a second layer of obligations under the AI Act. Changing this would mean that such AI systems would not automatically fall under the strictest requirements. Civil society is worried about leaving the technology without clear safeguards.
Political divisions
The EPP put forward the proposal, and it is dividing lawmakers. In an interview with EU Perspectives, MEP Leila Chaibi (The Left/FRA), shadow rapporteur for the Digital Omnibus in the IMCO committee, made clear that she opposes the change. “We are surely against. It’s a major point and one, among many others, of the reasons that made The Left vote against the EP position.” Ms Chaibi argues that such changes would weaken consumers’ safety, “notably in relation to toys and medical devices”.
This is not simplification, but complexification. Instead of one standard for high-risk AI, EPP now wants twelve.
— Kim van Sparrentak, MEP (Greens–EFA/NLD)
She also pointed to divisions between institutions and member states. According to Chaibi, “the Council, except for Germany, appears to be really reluctant on this issue”. Similar concerns were raised by Greens MEP Kim van Sparrentak (Greens–EFA/NLD). “I am still very concerned by the German Christian Democratic proposal to exempt medical devices, toys, smartwatches, and smart glasses from the AI Act,” she said. For Ms van Sparrentak, the proposal would not simplify the system. “This is not simplification, but complexification. Instead of one standard for high-risk AI, EPP now wants twelve.”
Double regulation fault line
For groups like DIGITALEUROPE, this overlap is a practical problem. The organisation argues that many companies, already regulated under sectoral frameworks, are caught in a “double or even triple layer of regulation”. “Overlapping, and potentially conflicting, documentation and conformity assessments risk delaying certification and slowing the deployment of innovative products,” the organisation states. Industry representatives call for AI requirements to be handled within existing sectoral frameworks rather than through the EU AI Act.
Civil society warns of loopholes
However, this is strongly contested by another group of actors. More than thirty organisations, including consumer groups, doctors, and digital rights advocates, warned against weakening the AI Act. Their concern is that sectoral laws do not address AI risks, such as opacity, bias, or autonomous decisions. In their words, the AI Act was designed “to regulate AI-specific risks which are not explicitly addressed in other sector-specific legislation.”
The health sector is one of the main examples. The organisations warn that medical devices and in vitro medical devices could be removed from the scope of the AI Act, even though such products increasingly use AI to influence diagnosis, treatment decisions, or patient monitoring. They also point to toys with AI and gaps in the protection of children.
Magdalena Maier, Legal and Advocacy Officer at CDT Europe, an NGO focused on digital rights, echoes these concerns. “Obligations and procedures, such as the reporting of serious incidents or dealing with products presenting a risk, have to specifically take into account risks to fundamental rights under the AI Act. This is not the case under sectoral legislation”, she told EU Perspectives. According to the expert, “the AI Act adds a fundamental rights-focused layer to its obligations and understanding of risks that is currently absent in the product safety field”.
You might be interested
Omnibus debate close to end
This debate is only one part of the broader Digital AI Omnibus negotiations. With trilogue negotiations expected to end on Tuesday, the final compromise will be voted on in June. The June vote will show whether Brussels can balance regulatory coherence with consumer protection. Whether the two can ever truly go hand in hand remains to be seen.
