EAB seminar plays down EU parliament’s biometrics ethics concerns

The European Union’s ethical concerns connected with deploying artificial intelligence (AI) and biometrics applications across its territory may be excessive.

The claims come from Emilio Mordini, an associate research fellow at the University of Haifa and Project Evaluator and Reviewer at the European Commission.

Speaking at an EAB session on Tuesday, Mordini first listed the prohibited AI practices contained in the upcoming AI Act, as well as the legislation’s applications that are deemed “high-risk.”

Public facial recognition and some other biometric systems are included in the current draft of the AI Act under this category, but some EU parliamentarians have since called for a revision to extend high-risk status to all systems receiving biometric data.

Mordini also suggests that while many in the EU parliament and the general public are concerned about facial recognition technology being used for surveillance and emotion recognition, he remains skeptical of doomsday scenarios connected with these tools.

In particular, Mordini mentions the “three gunmen problem,” as depicted in ‘The Good, the Bad and the Ugly,’ intended as “a confrontation in which there is no rational strategy that allows either party to win.”

According to the philosophy professor and psychiatrist, such a ‘Mexican standoff’ can only be resolved with each player anticipating the move of the other two opponents before deciding on his own action. AI cannot solve this problem.

“Machines, including Al, only know chronological time, which refers to the measurable course of events,” Mordini says. “But what really matters in life is timing.”

In other words, the professor argues that this limitation prevents Al from understanding and predicting individual behavior.

“Consequently, the concerns of the European Parliament are probably excessive because they attribute to Al a capability it does not possess and, in my opinion, will never be able to develop,” Mordini adds.

“Of course, working with big data, Al could predict the behavior of groups of people reasonably well. My suggestion is, therefore, to pay much more political and ethical attention to processes of categorization rather than to the unlikely use of Al to ‘read the minds’ of individuals.”

In contrast, those biometric applications that are banned or currently considered high-risk are actually feasible.

The EAB webinar comes days after the European Digital Rights (EDRi) published a new guide about remote biometric identification and how to legally regulate these technologies.

Article Topics

AI  |  AI Act  |  biometric identification  |  biometrics  |  EAB  |  ethics  |  facial recognition  |  legislation  |  regulation