Hungary's Fidesz releases AI-generated video depicting fictional wartime execution
Hungary's Fidesz party released an AI-generated election campaign video depicting a fictional wartime execution to discourage support for opposition Tisza Party, and disseminated deepfake videos of opposition leader Péter Magyar without clear labels, raising concerns about the use of synthetic media to manipulate democratic processes.
What happened
Fidesz published a 33-second AI-generated campaign video on Facebook depicting a blindfolded Hungarian soldier being shot, intercut with imagery of a crying child, implying that opposition Tisza Party would lead Hungary into the Ukraine war if elected.
Deepfake videos of opposition leader Péter Magyar allegedly smashing a cardboard cutout of Prime Minister Viktor Orbán were also disseminated through state-aligned media and social media groups including 'Harcosok Klubja' (Warriors' Club), typically without clear labels.
Reported impact
The blurring of reality and fiction threatens to dismantle a shared factual reality, leading to extreme political polarisation and a 'rally-around-the-flag' effect driven by manufactured fear.
The high production quality and lack of prominent watermarks misled many citizens, with government officials arguing the videos were 'obviously' AI or 'just memes' despite the deceptive presentation.
Index analysis
Government officials claimed the videos were not serious political communications but rather obvious synthetic content or memes, yet the lack of transparency and absent prominent watermarks enabled them to mislead audiences.
The incident occurred during a heated parliamentary election campaign, with Fidesz increasingly relying on AI tools to generate political content framing its opposition as dangerous or aligned with the European Union.
Governance controls
The incident highlights the urgent need for mandatory, non-removable watermarking on synthetic media and stricter liability for political parties using synthetic media to influence democratic outcomes.
The case serves as a test of the EU AI Act and election integrity standards in addressing the use of AI-generated content in political campaigns.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.