Meta declares that 'Lula's Army' channel does not violate its policies
Read more
Grupo Globo
g1.globo.com

Meta declares that 'Lula's Army' channel does not violate its policies

The company Meta declared this Friday, the 23rd, that the WhatsApp channel named 'Lula's Army' is not in disagreement with the guidelines established by the company.

In its attached technical declaration, Meta detailed that the analysis of the channel confirmed the absence of policy violation, also presenting the methodology used to identify infringing content. Furthermore, the response addressed data protection issues, providing information on available data, forwarding, and clarifying that there is no paid service on this interface.

This statement occurred in response to a request made by the president of the Superior Electoral Court (TSE), Kassio Nunes Marques, who granted a 24-hour deadline for Meta to send a comprehensive technical report on the 'Lula's Army' channel.

Nunes Marques' request was motivated by a requisition from the PL candidate's campaign for President, Flávio Bolsonaro, regarding the verification of the compliance plan signed by the company with the court. This plan describes the actions implemented by Meta to prevent and reduce risks to the integrity of the electoral process.

The judicial decision requires Meta to specify the technical criteria it uses to distinguish between 'legitimate mobilization' and 'potentially irregular coordinated behavior.' The minister responsible for the case emphasized that the focus of the evaluation lies in the 'confrontation between measures formally communicated to the Electoral Justice and the functioning of the questioned mechanism.'

Additionally, Meta informed this Friday that WhatsApp channels operate under different rules than those applied to Instagram and Facebook. The messaging app's terms of service explicitly prohibit the use of the platform for purposes such as 'coordinating harm, publishing falsehoods or distortions, impersonating someone, or sending illegal or unauthorized communications, such as mass messages, automated messages, or auto-dialing.'

Similar stories

Meta AI generates distinct results in requests involving Lula and Flávio Bolsonaro, says company
Read more
olhardigital.com.br

Meta AI generates distinct results in requests involving Lula and Flávio Bolsonaro, says company

Meta informed that it has implemented a temporary block on creating images of candidates through its artificial intelligence (AI) tools, aiming to comply with current electoral regulations in Brazil.

This stance was disclosed to Olhar Digital after tests indicated that Meta AI, accessible via WhatsApp, rejected requests related to Luiz Inácio Lula da Silva (PT), but allowed the generation of artwork containing the name of Flávio Bolsonaro (PL).

Complaints about the tool's behavior began appearing on social media on Tuesday, the 13th. During tests conducted by the report, commands mentioning Lula were blocked, while a similar request using Flávio Bolsonaro's name resulted in the production of an image of a green t-shirt with yellow inscriptions.

The tests were successful when the command asked for generation using only the candidate's first name from the PL. However, attempts to replace 'Flávio' with 'Bolsonaro' in the command were blocked. In the case of the current President of the Republic, it was not possible to obtain generation using 'Lula' even with his first names, such as 'Luiz Inácio'. The incorrect spelling of Lula's name, 'Luís Inácio', was also tested and this attempt was also blocked.

In response to Olhar Digital's inquiry, the company stated that it has taken measures to prevent the creation of candidate images, even when the commands are restricted to the first name or the electoral number. Meta stated: 'In compliance with Brazilian electoral rules, we have temporarily prohibited the recommendation of candidates or the generation of their images. We have taken additional measures so that commands using only the candidates' first names and their ballot numbers will also not produce images.'

After receiving Meta's statement, Olhar Digital made a new image request with 'Flávio', 'PL', and '22'. This time, the system temporarily allowed generation, but immediately blocked the attempt.

On Wednesday, the 14th, Lula's campaign filed new inquiries with the Superior Electoral Court (TSE) regarding the functioning of Meta's AI. The presidential coalition alleged that the tool demonstrates bias by restricting requests linked to the PT politician, while initially allowing the creation of images with references to Flávio Bolsonaro.

The campaign requested immediate actions to suspend the limitations imposed on electoral content and sought double compensation for the reach it considers to have lost. Furthermore, the coalition raised questions about other incidents on the platform, including issues with view metrics and the exclusion of profiles considered irregular.

In a previous statement before the Electoral Justice, Meta argued that a failure in view metrics was a global and temporary event, affecting more than one hundred countries, and that this did not impact the actual distribution of publications. The company also refuted the allegations of political favoritism made by the campaign.

Popular