AI-Generated Deepfake Audio Used to Influence Slovak Parliamentary Election (2023)

Incident Summary

In the days preceding Slovakia’s September 2023 parliamentary election, AI-generated deepfake audio recordings surfaced on social media during the legally mandated pre-election moratorium period, when candidates are prohibited from campaigning.^[1] The fabricated recordings purported to capture a phone conversation between Michal Simečka, leader of the pro-Western Progressive Slovakia party, and journalist Monika Tódová of Denník N, in which manipulated voices discussed strategies for rigging the election and purchasing votes from the Roma community.^[1][3]

The deepfake’s release coincided with a press release from Russia’s SVR intelligence agency making parallel claims about foreign interference in the election, which analysts interpreted as part of a coordinated influence operation.^[3][5] The incident is widely regarded as one of the first documented uses of AI-generated deepfake audio to interfere in a parliamentary election in a European Union member state.^[1][2]

Key Facts

• Target: Michal Simečka (Progressive Slovakia leader) and journalist Monika Tódová (Denník N)
• Method: AI-generated deepfake audio of a fabricated phone conversation
• Content: Fabricated discussion of vote-rigging and vote-buying from the Roma community
• Timing: Released during the 48-hour pre-election moratorium when candidates cannot publicly respond
• Context: Release coincided with a Russian SVR press release making parallel claims
• Election: Slovak parliamentary election held September 30, 2023

Threat Patterns Involved

Primary: Disinformation Campaigns — AI-generated deepfake audio was used as a tool in a broader information operation designed to undermine public trust in a democratic election and a specific political candidate.

Secondary: Deepfake Identity Hijacking — The identities of a political leader and a journalist were hijacked through AI voice synthesis to fabricate a conversation that never occurred.

Significance

1. First documented EU election deepfake interference. Widely regarded as one of the first cases of AI-generated deepfake audio used to interfere in a parliamentary election within a European Union member state.
2. Strategic timing during moratorium. The deepfake was released during the legally mandated campaign silence period, preventing the targeted candidate from publicly responding or rebutting the fabricated content before voters went to the polls.
3. Coordination with state intelligence. The coincidence of the deepfake release with an SVR press release making parallel claims suggests coordination between technical deepfake operations and state-level information operations.
4. Precedent for AI election threats. The incident has been widely cited in policy discussions about AI regulation, election integrity, and the need for rapid-response verification mechanisms during election periods.