This recent technology is gaining momentum in the territory. The so-called “intelligent” cameras, bbased on increasingly efficient algorithms, rely on software infinitely more powerful than our eyes and allow us to see what a human would have a hard time detecting. The Datakalab company has developed an algorithm capable of measuring the wearing of a mask in transport. “You see a free video stream of course, describes Xavier Fischer, the founder of Datakalab. In red, these are the people who wear the mask badly and in green, those who wear it well.”
According to him, everything is done to anonymize the people analyzed: “We don’t know who is who. Moreover we aggregate the data and we deliver it to our customers, therefore transport players, every twenty minutes and every hundred people minimum, to be certain that an individual who does not wear his mask is not verbalized via this data ” . He also sees a great statistical tool.
“When you go to a concert, there’s a person with a clicker in their hand to count the spectators. It’s the same here.”
Xavier Fischer, founder of Datakalabat franceinfo
Xavier Fischer lists possible applications. “If we take the example of cities, [il y a un intérêt] to understand flows. To understand where there are more or fewer people, where there is more or less traffic to manage traffic lights, he explains. Seeing the percentage of bikes, the percentage of cars, in the context of a mall, will optimize the number of cashiers in stores according to the traffic.”
Other companies go much further, with tools capable of spotting suspicious behavior, in stores or on the street. This software can detect attempted theft, illegal dumping of garbage or annoying parking. Valenciennes, Nice, Suresnes, Compiègnes… more and more cities are equipping themselves with these devices. And that is unacceptable for Arthur Messaud, of La Quadrature du Net, an association which campaigns against generalized surveillance: “We rather tend to reject all automated surveillance techniques, because the goal is going to be the same, he assures, it would be to analyze the population in a massive way as a whole to obtain a very, very detailed knowledge of it, which will allow the police to adapt their repressive activity.
“All of these automated monitoring systems are never imagined or thought to provide help or care to people.”
Arthur Messaud, from La Quadrature du Netat franceinfo
Arthur Messaud is formal, for him: “Whenever we talk about artificial intelligence, in practice, it is always to measure populations deemed dangerous or to repress them.”
The National Commission for Computing and Liberties (CNIL), the data protection policeman, is launching a major consultation on its site until March 11, and calls on the French as well as the companies concerned to give their opinion. This consultation does not concern “facial recognition/biometric recognition” which is already subject to a regulatory framework. Enough to fuel a possible legislative debate.