Questão de Inglês — compreensão de texto e estratégias de leitura

Police forces in several countries have started to use live facial recognition, a technology that compares the faces of passers-by with images stored in watch lists. Cameras mounted on vans or poles scan crowds in real time, and an alert is sent to officers whenever the software finds a probable match.

Advocates say the system helps locate fugitives and missing persons far faster than traditional methods. In some trials, wanted suspects were arrested within minutes of walking past a camera. Moreover, the technology does not get tired or distracted, unlike human officers watching hours of footage.

Civil liberties groups, on the other hand, warn that mass scanning turns everyone into a potential suspect. They also raise concerns about accuracy. Earlier generations of algorithms were found to produce more false matches for women and people with darker skin, partly because they had been trained on unrepresentative data sets. Developers claim that recent versions have narrowed this gap, but independent audits are not always available to confirm it.

A false match is not merely a technical glitch. For the person wrongly stopped, it may mean being questioned in public, searched or even detained. Critics argue that the burden of proving innocence should never fall on someone simply because a machine flagged their face.

Some cities have banned the use of facial recognition by public agencies altogether, while others allow it under strict conditions, such as judicial authorisation, clear limits on how long images can be stored and regular reporting on results. Lawmakers in many places, however, have yet to set any specific rules, leaving decisions largely in the hands of police leaders.