We submitted a public comment to the Meta Oversight Board in response to case 2026‑029‑FB‑UA, which concerns an apparently AI-generated video depicting Hungarian politician Péter Magyar, posted on Facebook and viewed more than 100,000 times before the 12 April 2026 parliamentary elections.
The Oversight Board is examining Meta’s decision that the content did not violate its Misinformation Community Standard and did not require an AI label, in part because it was posted “well in advance” of the elections and appeared intended for comedic effect.
In our submission, we argue that this framing reflects an unduly narrow understanding of electoral risk. We also emphasise that AI-generated impersonations of political candidates raise distinct risks beyond ordinary satire as they shape voters’ intuitive judgments of competence, temperament and trustworthiness. The submission documents the broader context of AI-generated content, coordinated inauthentic behaviour and hostile foreign influence during the 2026 Hungarian elections, and highlights weaknesses in Meta’s triage and escalation systems for AI-generated political content in high-risk electoral settings. We conclude with five concrete recommendations, including prioritising network-level threats, ensuring adequate local-language human moderation, and deepening structured cooperation with EU institutions and national regulators.
Our full comment is available here.