Lawmakers draft rules for high-risk AI in public services

A parliamentary working group published a draft for high-risk artificial intelligence used by government agencies. The text would ban certain biometric uses in public spaces and require agencies to keep model cards and decision logs.

Technology vendors asked for a single national standard rather than a patchwork of city rules. Civil liberties groups argued that impact assessments must be published, not filed internally.

A consultation period runs for eight weeks. A final bill is expected before the winter recess.