With Brazil’s national election scheduled for October 4, the Tribunal Superior Eleitoral (TSE) is finalizing a definition of deepfakes and setting new rules for AI-driven campaign material. This framework aims to address complex issues raised by synthetic content in politics, including transparency and algorithmic recommendations, amid ongoing debates over how to manage emerging AI technologies in elections.
- TSE to define deepfakes and enforce disclosure rules on AI-generated election content.
- New restrictions include bans on AI-driven recommendation systems and election blackout periods.
- Debate continues over synthetic content transparency amid Bolsonaro’s AI avatar controversy.
What happened
Brazil’s Tribunal Superior Eleitoral (TSE) has been deliberating to define what qualifies as a deepfake in the context of the upcoming October 4 national election. The court is preparing to implement a broader regulatory framework on AI-generated campaign materials, which goes beyond simply banning deceptive footage. These new rules include mandatory labeling of AI-synthetic content, prohibiting platforms from recommending specific candidates using algorithms, and enforcing a blackout period on publishing or promoting synthetic political content immediately before and after the election.
The urgency of these new regulations stems from recent incidents such as former President Jair Bolsonaro’s appearance as an AI-generated avatar during his son’s campaign launch, despite his home confinement and communication restrictions. Although the avatar was explicitly disclosed as a simulation, the case has sparked legal debates about the boundaries between permissible synthetic content and prohibited disinformation, as well as the interpretation of existing electoral laws.
Why it matters
Brazil’s approach addresses one of the most challenging dimensions of AI and politics: how to balance transparency against potential misinformation and manipulation. While disclosure rules are designed to alert voters when content is synthetic, the Bolsonaro case illustrates that labeled AI content can still be politically influential, blurring the line between clarity and deception. Furthermore, a politician legally barred from public statements can remain present and persuasive through AI-rendered representations, complicating regulatory and ethical frameworks.
The restrictions on AI-driven candidate recommendation systems also extend regulatory attention from content creation to content distribution, challenging platforms' algorithmic mechanics that shape voter exposure. Brazil’s courts have a history of assertive intervention in regulating political speech and online content, positioning the country at the forefront of experiments with AI election oversight. The decisions made now will have ripple effects potentially influencing global regulatory approaches amid growing concerns over AI’s role in democratic processes.
What to watch next
The final TSE ruling will clarify the boundaries of acceptable AI use in elections and set a precedent for enforcement in a narrow window before voting day. Campaigns, social media platforms, and regulators must adapt quickly to ensure compliance with disclosure requirements, blackout periods, and restrictions on algorithmic recommendation. Monitoring how these rules affect political communication and voter influence will be critical in assessing Brazil’s regulatory model's efficacy and fairness.
International observers and policymakers will also closely analyze Brazil’s evolving electoral policies as a possible template for future AI governance in elections globally. Key issues to watch will include how the court reconciles the tensions between free speech, candidate restrictions, and technological innovation, and how enforcement tackles the complex challenges posed by synthetic media at scale.