Thousands of London Underground passengers have been observed and analyzed without their consent for at least a year through artificial intelligence (AI) systems powered by surveillance cameras. The facial recognition deployed makes it possible to detect the gender, age or potential emotions of the user, with the suggested purpose of using all this information to design a type of personalized advertising. The technology has been developed by the Amazon company.
The news has been published by the technology information page Wired, who has had access to the documents obtained by the organization for the defense of civil rights and liberties Big Brother Watch (Big Brother Watch). Through a request made under the protection of the Freedom of Information Law, the NGO has obtained all the documentation related to the tests carried out – at least since April 2023 – in eight stations of the suburban network of the British capital. The most important are Euston, Waterloo or Manchester Picadilly. The tests have been carried out under the apparent supervision of Network Rail, the public company that manages and owns most of the UK’s railway network.
The initial purpose of the project is to develop a system for detecting possible accidents, security risks or criminal acts or vandalism. The technology deployed at the stations uses recognition of objects detected in the cameras, through machine learning or machine learning, to be able to warn, for example, of users jumping onto the tracks, of an excess of passengers on the platforms or of attitudes that alter the order of the network, such as people running, shouting or smoking. Through wireless sensors, the system also detects slippery floors or possible flooding.
But the most alarming thing for the organization that has obtained the information is thatSome elements of the tests deployed focus on the “demographic traits” of the passengers. The images obtained by the cameras, according to the documents made public, can produce “a statistical analysis of the age and gender range” of the users, as well as “analyze some emotions” to detect whether they are sad, happy or irritated.
“The deployment and normalization of surveillance through AI in public spaces, without any prior consultation or debate, is a very worrying step,” has told Wired Jake Hurfurt, the director of research at the ONg.
All the images, according to the information, are captured when passengers cross a virtual line located just in front of the entrance barriers to the stations. Data is sent to the system Rekogntion from Amazon, trained to perform facial and object analysis. It allows measuring the degree of “satisfaction” of users and, according to the revealed documentation, “the data could be used to maximize advertising and retail sales.”
“We take the security of the railway network extremely seriously and use a variety of technological advances, deployed at all of our stations, to protect passengers, employees and the infrastructure itself from possible threats or criminal acts,” he assured. a spokesperson for Network Rail, who did not want to comment on the specific use of AI in the underground. “When we deploy new technology, we do so jointly with the police and security services, so that the measures are proportionate and in accordance with the law.” “, has added.
The system deployed in the London Underground does not use the currently most questioned facial recognition technology, which links detected users with identities already classified in a data storage.
You can follow The USA Print in Facebook and x or sign up here to receive our weekly newsletter.
