Christiane Wendehorst

Biometric techniques raise specific ethical issues, as an individual cannot easily change biometric features, and as they tend to intrude into the human body and ultimately the human self. The talk will analyse how the Proposal for an AI Act deals with these techniques.

Biometric identification together with biometric categorisation, behavioural detection, emotion recognition, brain-computer-interfaces (BCIs), and similar techniques are being used to an increasing extent by public and private bodies. They serve a broad variety of purposes, ranging from healthcare to law enforcement and border control to warfare, and are deployed in public as well as in private spaces. Beyond traditional biometric techniques such as fingerprint or facial recognition, ‘second generation biometrics’ include, e.g., analysis of keystroke or mouse dynamics, gesture dynamics, signature dynamics, as well as voice and gait features. Major trends are the increasing use of ‘weak’ and ‘soft’ biometrics alongside ‘strong’ biometrics, focussing on a variety of patterns of a more behavioural kind, and the development towards multimodal biometrics. The Proposal for an Artificial Intelligence Act (AIA) of 21 April 2021 addresses such techniques in various ways, as do other instruments, both existing and in the pipeline. However, the question arises whether existing and proposed legislation adequately addresses ethical and fundamental rights issues raised.