The advancement of digital technologies and the growing use of Artificial Intelligence (AI) in political campaigns have led the Electoral Court to update the rules for the election period. The new resolution from the Superior Electoral Court (TSE) establishes criteria for the use of content produced or manipulated by AI, in addition to outlining restrictions, responsibilities of digital platforms, and possible penalties in case of irregularities.
In an interview with Folha do Litoral News, lawyer Waldir Franco Felix Junior, a specialist in Electoral Law and Constitutional Law from the Franco Félix law firm, analyzes the main points of the regulation, the risks of misinformation, and the challenges to ensuring balance and transparency in the upcoming elections.
Check out the interview:
Folha do Litoral News: What motivated the Superior Electoral Court to establish new rules for the use of artificial intelligence in election campaigns?
Attorney: The main reason was the increase in such mechanisms since the last elections, as well as a widespread perception that the old rules were not perfectly suited to the practices observed during the campaign.
Folha do Litoral News: What are the main points of the resolution that deals with the use of AI during the election period?
Attorney: In my opinion, there are two. The first is the possibility of using AI in campaign materials, provided that no other illegal activity occurs (such as anonymity, fake news, etc.), as long as the content expressly indicates that it was produced or manipulated with AI. The second, equally relevant, is that its use is absolutely prohibited in the 72 hours preceding the election and in the 24 hours following it.
Folha do Litoral News: How do you assess the requirement that content produced or manipulated by artificial intelligence be identified as such? Could this reduce misinformation?
Attorney: Certainly. In the difficult attempt to combat misinformation, this measure is essential.
Folha do Litoral News: In practice, how will the obligation of digital platforms to remove illegal content work, even without a court order?
Attorney: This measure is perhaps one of the points that requires the most explanation. By reversing the burden of proof and allowing/giving the platforms the power to remove content spontaneously, they will receive enormous power. This does not mean, however, that one can infer an effective interest in proceeding in this way, given that, historically, these platforms are not inclined to this type of action. Therefore, even if in a somewhat uncertain way, it will be up to them to suspend content that proves to be spreading misinformation, fake news, or other possible electoral offenses, only after guaranteeing the right of its creator/reproducer to defend the legality of the respective content.
Folha do Litoral News: Does this ruling by the TSE (Superior Electoral Court) impose greater responsibility on social media networks? Is there a risk of excess or preventive censorship?
Attorney: Yes. As previously stated, this greater competence brings with it a risk of excesses, which ultimately lead to the risk of censorship. This is the central point of the analysis that should be carried out during the electoral process.
Folha do Litoral News: What is considered "illicit content" within this new context involving artificial intelligence?
Attorney: Based on my previous answer, any content that does not expressly identify itself as AI-generated, or content created between 72 hours before and 24 hours after the election, will be considered illegal.
Folha do Litoral News: Regarding so-called "synthetic content," what exactly does the standard mean by this term?
Attorney: Generally, it refers to audio, image, video, or text material that has been wholly or partially generated or artificially manipulated by digital technology (especially AI), appearing to be something that did not actually occur.
Folha do Litoral News: What types of penalties could candidates and parties face if they violate the new rules?
Attorney: The main consequence is the removal of the content. Furthermore, it is crucial to highlight the possibility of investigating abusive conduct, especially in the form of misuse of media, which could lead to the revocation of registration or diplomas.
Folha do Litoral News: How can voters identify if they are facing content manipulated by AI?
Attorney: Voters can identify potentially AI-manipulated content by observing certain common signs in images, videos, and audio. In many cases, small visual inconsistencies appear, such as hands with an unusual number of fingers, distorted text on signs or clothing, blurred edges on the face, or subtle changes in appearance throughout the video. Manipulated recordings may also exhibit synchronization problems between speech and lip movement, unnatural facial expressions, or a voice with an artificial intonation. These details don't always appear, but they can indicate that the content was digitally generated or altered.
Furthermore, it is important to assess the context and origin of the information. Content that causes surprise, outrage, or immediate scandal, especially when it appears without a reliable source, deserves verification before being shared. Voters can check if the information was disseminated by recognized media outlets, official profiles, or other independent sources.
