Already today, the roads are being monitored automatically in some German states. However, only the respective number plates are recorded and compared with a database. This is only allowed if there is a specific reason. However, the system is error-prone anyway, because the license plate can be forged on the one hand. On the other hand, the license plate says nothing about who actually controls the car. In New York, therefore, on the Robert F. Kennedy Bridge last year, a new software was tested, which should identify the actual driver with automated facial recognition. However, as the "Wall Street Journal" now reports, the technology proved to be not very practical.
The tests should be continued in other places
Specifically, none of the subjects could be recognized correctly. The reason was probably the high speed of the vehicles and the general confusion of traffic on the bridge. For in advance, the "Oak Ridge National Laboratory" had also tested the software and achieved a marksmanship of around eighty percent – but only at significantly reduced speed. The authorities in New York do not want to be impressed by the setback in real conditions. Instead, a spokesperson said the tests should continue at various locations within the metropolis. However, it may take several years before the technology can actually be used by the security forces.
Even with pedestrians, the technology does not work properly
Especially since the automated face recognition of pedestrians does not work properly. For example, British police around the Champions League final in Cardiff in 2017 faced an extremely high number of false positives . In other words, the software has struck 2,297 times with perfectly respectable citizens. In addition, the automated systems generally have the problem that they recognize certain people better than others. Thus, the hit rate for white men is comparatively high. Black people or women, on the other hand, have greater problems with technology. In China, therefore, is already working on a technology that will recognize people by their gait . But privacy advocates warn against all these automated detection systems.