European organizations can run into problems when assessing whether artificial intelligence systems unintentionally discriminate based on ethnicity. Such sensitive data use is a violation of the EU General Data Protection Regulation, but the European Commission's proposed AI Act includes a provision allowing organizations to use sensitive data for AI audits. Researchers Marvin van Bekkum and Frederik Zuiderveen Borgesius summarized their findings from a paper that explores the conflict between regulations and a potential resolution.
Full Story