Senate hires facial recognition software accused of collecting data without authorization

The Federal Senate acquired the facial recognition program Clearview AIdeveloped by the American company of the same name, which faces accusations of improper collection of more than 50 billion images on the internet. The contract, signed in November, provides for the use of the tool by agents of the Federal Legislative Police.

Clearview AI

The software works by comparing scanned facial images with photographs stored in a vast database, allowing the identification of individuals and the display of information such as names and social media profiles. Despite its technology, Clearview AI has been the target of criticism and lawsuits around the world due to concerns about privacy and ethics in the use of data.

System Use and Limitations

According to the contract, Clearview AI will not be used as a real-time surveillance tool, but rather to support investigations after crimes or infractions have occurred. The Legislative Police hopes that the solution will speed up image analysis, currently considered a “complex and time-consuming” task. However, the Senate stated that the software will not completely replace human work, acting only as a complementary resource.

The license purchased for R$87,680 allows unlimited queries to the Clearview AI database, including training and support. The contract will be valid until November 2025.

Ukraine uses Clearview AI

In 2022, Ukraine revealed that it began using Clearview AI software to help discover Russian attackers, combat disinformation and identify the dead, information highlighted by the company on its official website.

Controversial

With more than 50 billion images and an accuracy rate greater than 99%, according to the company, Clearview AI has been the target of several international lawsuits. The accusations point to privacy violations due to the collection of images without consent. In the United States, a class action lawsuit seeks compensation for the misuse of photographs, while giants like Google and Meta are putting pressure on Clearview AI to stop collecting data from their platforms.

Em 2020, o The New York Times revealed that the The company used photos available on social networks, Google search, YouTube videos and other platforms without users’ consent to train its facial recognition system.

Clearview AI this year suggested that compensation for people affected by the hacking of its search engine be paid with shares. The company allocated around 23% of its shares, equivalent to US$52 million, for this purpose.

Hoan Ton-That, co-founder and executive director of Clearview AI, has made numerous public statements over the years reinforcing his company’s working model. In one of the statements, the executive stated the following:

“Our solutions enable agencies to gain intelligence and disrupt crime by uncovering clues, insights and relationships to help investigators solve simple and complex crimes, increase public and police safety, and keep our communities and families safer.

In Brazil, the situation also raises concern. To Lei General Data Protection (LGPD) requires that the use of personal data by companies is authorized by the holder of the information. Without this consent, there is the possibility of legal violation and resulting penalties for Clearview AI.

Senate Response

In a note to G1, the Senate defended the acquisition of the program, stating that it does not violate Brazilian legislation. According to Casa, the contract includes clauses that guarantee compliance with the LGPD. The agency also highlighted that other public institutions, such as the Federal Police and the Ministry of Justice, already use similar tools.

“The provision of a clause that guarantees the correct use of information, more than a legal requirement, is the balance between police activity and the guarantee of privacy”stated the Senate in an official statement.