Hungarian Election Campaign Marred by AI-Generated Disinformation as Orbán Seeks Fourth Term
Key Takeaways
- ▸AI-generated videos depicting execution and fabricated political conversations were weaponized in Hungary's 2024 election campaign without proper disclosure by ruling party affiliates
- ▸The disinformation campaign targeted opposition leader Péter Magyar with false claims he would involve Hungary in Ukraine's war if elected—allegations his party's manifesto explicitly refutes
- ▸While some AI content was acknowledged as synthetic, creators did not consistently disclose the technology's use, allowing fake videos to spread millions of times across social media before fact-checking
Summary
As Hungary prepares for pivotal elections on April 12, Prime Minister Viktor Orbán's Fidesz party and allied groups have deployed AI-generated videos as campaign tools to attack opposition rival Péter Magyar. The most prominent example is a disturbing fake video depicting a Hungarian soldier's execution, posted to Fidesz social media in February, which falsely suggests Magyar's election would lead to Hungarian involvement in Ukraine's war. A second AI video, viewed 3.7 million times, fabricates a conversation between European Commission President Ursula von der Leyen and Magyar discussing funding for Ukraine, initially posted without disclosure of its synthetic nature by the pro-Fidesz National Resistance Movement (NEM). While Fidesz acknowledges some AI content is fake, the party has not responded to questions about the creation and posting of these manipulated videos, instead using them to advance unsubstantiated claims about Magyar's pro-war intentions—allegations the opposition candidate and his Tisza party have firmly rejected. Despite the campaign's heavy reliance on generative AI disinformation, Magyar currently leads in most opinion polls, suggesting the strategy has had limited electoral impact.
- The unprecedented scale of generative AI use in Hungarian political messaging marks a significant escalation in election-era disinformation tactics, though initial evidence suggests limited voter impact
