New facial recognition ethics tool being tested in the Netherlands

A reality-check pilot project has been launched in the Netherlands to see if a proposed governance framework can mitigate the risks inherent in facial recognition surveillance.

The project is introduced and described in a new white paper published by the public-private World Economic Forum.

The result is not suggested as a standard for ethical use of biometric surveillance but instead as a “common set of proposed principles” for the same. It also contains a self-assessment tool for police agencies to check how closely they hew to the principles.

AI ethics tools are becoming more numerous of late, a natural (and welcome) development after a prolonged period of conceptual thinking.

Among the white paper’s authors is Marjolein Smit-Arnold Bik, head of special police operations in the Netherlands. Other writers include Cyril Gout, director of operational support and analysis for Interpol, and Irakli Beridze, leader of an AI center within the United Nations’ Interregional Crime and Justice Institute. The project community also includes NIST Biometric Standards and Testing Lead Patrick Grother.

Both the Netherlands police and Interpol investigators are experienced in using facial recognition systems for identification and authentication, according to the document.

They both also employ facial examiners, who are considered experts in biometrics analysis. These examiners work autonomously from crime investigators, using probe images and databases.

As the pilot project moves forward, law enforcement agents will be applying lengthy, enumerated principles with top-level headings like respect for human and fundamental rights, necessary and proportional use, transparency.

Other headings may not be as familiar to some governments still wading into operation standards, such as use of probe images and reference databases, image and metadata integrity and risk-mitigation strategies.

Some of the same headings, naturally, show up in the facial recognition ethics tool as well. The tool is designed to make sure policy makers and operational managers do not lose track of their duty to protect and serve their constituents.

During the pilot, the agencies will push the governance framework around to see if it is realistic and effective. Findings will be used to refine the framework. No date has been set for completion.

Article Topics

biometrics  |  ethics  |  facial recognition  |  Netherlands  |  pilot project  |  police  |  regulation  |  surveillance  |  World Economic Forum