Brazilians are winning court claims for AI failures, including on facial recognition

Brazil could be one of the best countries to claim compensation for failures in the use of artificial intelligence.

That’s the takeaway from a survey by law firm TozziniFreire Advogados, which found that such claims were upheld on appeal in 64 percent of these claims. Since 2010, when the term was first mentioned, the survey found that 140 filed lawsuits have sought compensation.

These 140 decisions had AI as the direct cause of action, while 711 rulings brought up AI at some stage of the process. Those with the most cases on AI were those of São Paulo (61 cases) and Rio Grande do Sul (49 cases).

TozziniFreira litigation partner Sofia Kilmar explained that most cases involved financial fraud prevention, AI for facial recognition, and removal of products violating marketplace terms of use. Previously, AI was only mentioned in cases relating to how internet content was organized, but from 2016 onward topics began to diversify. There were 50 cases that questioned search processes and algorithms used by technology platforms, according to Kilmar.

Kilmar said that the courts made no distinction between algorithm and AI in lawsuits, and that this is intentional in order to broaden out the analysis of the level of risk posed by the technology. Algorithms, generative AI, facial recognition, deepfakes and forms of automation are dealt with the same.

Bank fraud and AI were the main issue in 49 cases, with a legal precedent set by the Court of Justice of São Paulo (TJSP) that concluded a bank possesses sufficient technological knowledge to prevent fraudulent practices “including using artificial intelligence to detect consumer profiles and the transactions conducted.” Damages of R$5,000 were awarded to the plaintiffs.

Facial recognition was first mentioned in 2021 before appearing much more frequently, with problems related to recognizing black faces and lack of transparency in relation to use of sensitive data and storage. For example, there was a case involving a rideshare driver who was removed from the platform as the AI system did not recognize their face as the person registered. It’s observed that there is no specific legislation on the issue so the majority of civil cases rely on the Consumer Defense Code. For much more legal detail, Valor International has the laws.

According to Opice Blum partner Camillia Jimene, the biggest challenge legislatures and judiciaries will face worldwide will be in relation to intellectual property and AI.

Article Topics

accuracy  |  AI  |  biometrics  |  Brazil  |  facial recognition  |  lawsuits