Associated Incidents
The use of tools based on artificial intelligence (AI) in Brazil has become increasingly common in several areas, including the social security system. Meu INSS, an application created by the National Institute of Social Security (INSS), was implemented in 2018 with the aim of streamlining the process of granting benefits, such as retirement and sickness benefits, through automated analysis. Recently, the INSS began using AI in the process. However, the system has caused problems for many Brazilians, especially rural workers and those with little familiarity with technology, as shown in a report by Rest of World. The American non-profit publication reported the case of Josélia de Brito, a former rural worker from the Brazilian northeast who experienced first-hand the challenges imposed by the AI system. When she applied for retirement through the app, Josélia had her request automatically denied due to an identification error, and was registered as a man in the system. This type of error has become common for workers with little access to technological resources, who face difficulties in correcting inconsistencies, since the system does not offer effective review alternatives. The implementation of AI in Meu INSS --------------------------------- - Meu INSS was developed with the aim of reducing bureaucracy and speeding up the granting of benefits. - Since its introduction, the tool has been used to process a large volume of requests, with the government's goal that by 2025, 55% of requests will be analyzed automatically. - In many cases, the system has managed to speed up the processing of simple requests. - However, this has also led to the automatic rejection of applications with minor errors or inconsistencies, which particularly affects those in remote areas of the country.
- The introduction of AI into the process was initially seen as a positive innovation, with the promise of making the system more agile.
- However, as experts point out, the use of technology without critical analysis and without adequate training of citizens has resulted in digital exclusion.
- Josélia de Brito, for example, who suffers from chronic illnesses and needs social security benefits, reported her frustration at seeing her applications repeatedly rejected, even though she had all the necessary documentation.
- For her, the process of applying for benefits became humiliating and each error in the system seemed to result in a denial without clear explanations.
The impact on rural workers
The implementation of AI in Meu INSS has proven to be especially problematic for rural workers, who often do not have access to quality internet and face difficulties with digital literacy.
Data indicate that the illiteracy rate in rural areas in Brazil is three times higher than in urban areas, which increases the difficulties these workers have when interacting with digital systems. In addition, many complex cases, such as those involving rural properties that are difficult to document, end up being rejected by AI, which is unable to process these requests accurately.
Experts say that, despite efforts to streamline the process, automation is not prepared to deal with the particularities of rural work. Edjane Rodrigues, secretary of the National Confederation of Agricultural Workers, highlighted to Rest of World that these cases should be analyzed by human beings, since the social legislation provides for a series of nuances that AI cannot interpret correctly.
Can technology be a solution or an obstacle?
Although the Brazilian government has invested millions to improve the system, with the aim of making services more efficient, the growing dependence on artificial intelligence has also raised concerns. In early 2025, the company responsible for creating the app, Dataprev, announced an investment of approximately 10.5 million dollars to improve the data analysis and fraud detection capabilities of Meu INSS.
Despite criticism, Meu INSS continues to be widely used in Brazil, with 84 million monthly accesses. However, experts warn that increased reliance on technology could result in inefficiency and exclusion for the most vulnerable, who lack the resources or knowledge to navigate automated systems. In a scenario where the digitalization of public functions is seen as a solution to increase efficiency, there is a need for a balance between automation and human analysis. This is especially important in cases where the benefit granting process involves elements that cannot be reduced to simple algorithmic criteria. Without this balance, innovation can become an obstacle for the citizens who need assistance the most.