John Lewis, an 80-year-old African American deputy from Georgia, one of the icons of the battle for civil rights in America. But when, two years ago, the California Civil Rights League organized an experiment to demonstrate the unreliability, especially against ethnic minorities, of the facial recognition system, his face was exchanged by Rekgnition, the artificial intelligence of Amazon recognition for a criminal. The same thing happened with 27 other MPs whose faces had been compared with tens of thousands of mug shots from the digital archives of the Californian police: the clearest demonstration that these systems, in addition to extending the powers of the police in an extraordinary way, are unreliable and make many mistakes, especially in the recognition of black people.
For years, the leaders of the progressive movements have been fighting, not only to limit the harshness of the police in patrols or when making arrests, for a regulation of increasingly powerful technologies that, made available by the Silicon Valley companies to law enforcement agencies, they risk reproducing in the United States a surveillance society similar to what has already become a tragic reality in China. Experiments such as that of 2018 have caused a stir, prompting the Washington Parliament to carry out various fact-finding investigations, but laws have not been passed so far. Meanwhile, the collaboration of digital companies with many American police has grown. We also saw it in the days of the great American demonstrations: artificial intelligence increasingly used for facial recognition, cameras everywhere, demonstrators watched from above with drones and also by aircraft of the drug departments, the National Guard, the FBI and the police of border. The impact force conquered by Black Lives Matter at this stage has, however, enabled these movements to obtain for the first time an important result in comparison with the giants of technology. As for a domino effect, first IBM, then Amazon and finally also Microsoft have decided not to grant the use of their facial recognition technologies to the police of America anymore. Battle won? Not exactly. Firstly, the supplies were not canceled, but only suspended for a year. Probably waiting for a law that regulates the matter. And then in the field of recognition the large groups of Silicon Valley are not the most precious in the eyes of the police: Microsoft has a marginal role, while Amazon has relationships with law enforcement mainly through Ring, the surveillance system of individual homes or entire neighborhoods managed in collaboration with law enforcement agencies. A system based on cameras and not on facial recognition: therefore, not subject to the restrictions announced by the company. And in any case, Nec, Idemia and Clearview AI provide the facial security systems to the agents: three companies that have no intention of interrupting collaboration with law enforcement agencies.
So the undoubted victory of the image and also the change of climate, but on the practical results it is better to be cautious. Also because the police bodies, opposing security needs (understandable, especially in an anti-terrorist perspective), do not say which control tools they use (certainly not only facial recognition) nor how deeply they intend to go. Here the problem is not the pursuit of the tactics of the agents of the various states but the federal legislative framework. Congress regulatory intervention is needed if we want to prevent surveillance systems such as China from spreading in the United States, in the wake of the alarm for the unrest, also connected to mechanisms that reward or punish citizens based on their social behavior.
So far it was the giants of Silicon Valley who resisted in fear that politics would ban the development of some potentially problematic or even dangerous, but also promising and profitable technologies.
© RESERVED REPRODUCTION