The planned European AI law does not go far enough for data protectionists. The European Data Protection Authority (EDPB) and the European Data Protection Supervisor have published a joint statement on the European Commission’s plans. They praise the fact that the use of AI should be regulated, but call for stricter rules for certain areas.
The draft of the European Commission provides for a step model for assessing the risk of AI systems. Among other things, biometric monitoring using AI and social scoring systems would be classified as high-risk.
Prohibitions of risky applications
The data protection authorities are calling for a general ban on AI for the automated recognition of human characteristics in public spaces. Such systems would mean “the end of anonymity in these places” and restrict basic rights.
When it comes to social scoring, the authorities only consider medical use to be justified. Otherwise they are calling for a ban in this area as well. Intelligent systems that are supposed to evoke certain emotions in people would therefore violate basic values of the EU and could lead to discrimination.
Systems that assign people to specific groups with the help of biometric data should also be prohibited.
Clearer statements required
The authorities lack a clear statement that the data protection regulations also apply to the processing of data by AI. They demand that compliance with the regulations become a condition for entering the EU market.
The EDPB and the Data Protection Officer also request that concepts for international law enforcement cooperation be added.
The Commission’s draft will next be discussed in the European Council of Ministers and the Parliament. So there is still room for changes and possible tightening.
