Ban the use of facial recognition in publicly accessible spaces and classify artificial intelligence (AI) emotion recognition systems as high-risk systems. These are some of the main recommendations made by the European Consumer Association (BEUC) regarding the European law on AI, the first regulation in the world on the subject that aims to set the boundaries in the use of AI-based systems, ensuring a balance between innovation and respect for human rights.
The measure, proposed by the European Commission in April 2021, is now the subject of negotiations between the EU institutions, after the two co-legislators, the Council and the European Parliament, approved their respective negotiating positions. A crucial moment, according to Beuc, to improve the text, strengthening consumer protection since, it is noted, once the law is approved, the Member States will not be able to further regulate this sector.
One of the most controversial issues raised by the association concerns the classification of high-risk AI systems.
In essence, BEUC recommends returning to the original formulation of the European Commission, since the changes introduced by the Council and the EU Parliament would allow developers to subjectively assess the level of risk of their systems. A gap that could compromise the entire legislation, Beuc argues, stressing the need for an objective, coherent and legally certain process to identify "high-risk" AI systems in the AI law.
Another point on which the association focuses is generative AI, a complex topic that arrived when the proposal of the EU executive had already been presented. In this regard, specific requirements are called for for all AI systems, for generative AI systems, not only for high-risk generative AI systems so as to ensure a thorough assessment of such systems before they are placed on the market. Finally, among the priorities highlighted is the need to ensure that consumers are adequately protected when they are subject to an AI-system, for example by providing for the right to lodge a complaint if harmed by a high-risk AI-system or the right to be informed of the possible use of an AI-system.
All rights reserved © Copyright ANSA