Lien copié

Reconnaissance biométrique : quelle régulation ?

Regulating Biometrics : Global approaches and urgent questions : ce rapport  propose, à travers 8 témoignages, un panorama des différents types de régulation des technologies de reconnaissance biométrique. Ces technologies, qui connaissent une expansion rapide à travers le monde, recouvrent (1) la reconnaissance faciale, de la voix, du comportement ou de données corporelles. L’étude réalisée par de New-York University, NYU, aux Etats-Unis, permet de mieux appréhender, dans le cadre de différents contextes nationaux,  les différentes manières d’utiliser la loi comme un outil pour développer, ancrer ou limiter leurs utilisations. Elle s’accompagne de réflexions sur l’inéluctabilité du recours à ces technologies, les  performances réelles, les justifications et les différentes modalités contemporaines de leur mise en oeuvre, le cas échéant en marge de toute réglementation. “The last few years mark a critical juncture, perhaps even a turning point, in the trajectory of continued biometric expansion. Civil-society advocates have challenged the foundational arguments made by companies and governments that produce and promote these technologies, highlighting the tangible harms caused by their use » estime Amba Kak, éditrice du rapport. 

Amba Kak, ed., “Regulating Biometrics: Global Approaches and Urgent Questions” AI Now Institute, September 1 2020,


1- Que recouvre la reconnaissance biométrique ? “Although the terminology varies, we use the phrase biometric recognition technologies to describe systems that “fix” official identities to bodily, physiological, or behavioral traits, or tracked. While fingerprints have the longest history as a marker of identity and continue to be used in a number of applications across the world, other bodily markers like face, voice, and iris or retina are proliferating, with significant research exploring their potential large-scale application. Emerging areas of interest in this field include using behavioral biometrics like gait (i.e., how a person walks), keyboard keystroke patterns, and multimodal combinations of biometrics to identify and potentially make inferences about individuals. Beyond identifying people, these systems increasingly claim to be able to infer demographic characteristics, emotional states, and personality traits from bodily data. (This practice is sometimes referred to as “soft biometrics”5 in technical literature». P. 6. 

0 commentaire


Your email address will not be published. Required fields are marked *

Champs obligatoires*