Which YouTube deepfakes from 2023–2025 were tied to elections or major political events?

1. Slovakia’s election‑timed audio that went viral — the poster child for pre‑election manipulation

In September 2023 a fabricated audio clip attributed to Michal Šimečka, leader of the pro‑Western Progressive Slovakia party, circulated claiming he discussed plans to rig the parliamentary election, and researchers and watchdogs flagged it as a deepfake used to seed doubts about electoral integrity ^{[1] [4] [5]}.

2. “Hillary Clinton endorses DeSantis” — a cross‑platform video that reached YouTube

On 11 April 2023 a manipulated video showing Hillary Clinton apparently endorsing Florida governor Ron DeSantis for president spread across social networks and was reported to have been shared on platforms including YouTube, Instagram and X, becoming an early and widely cited example of AI‑generated political content in the U.S. context ^[2].

3. India’s 2024 election: celebrity deepfakes on WhatsApp and YouTube

Researchers and the Brennan Center documented AI‑generated videos and images in India’s 2024 general election that depicted celebrities criticizing Prime Minister Narendra Modi or endorsing opposition parties; those artifacts went viral on encrypted messaging apps and on public platforms such as YouTube, illustrating how synthetic media crossed technical and social boundaries to influence voter conversations ^[3].

4. Turkey and other international cases where video clips were weaponized around votes

Analysts recorded manipulated visual material in Turkey’s 2023 contest and other competitive races — including videos circulated to link opposition candidates to terrorist groups — with reported use by campaigns and partisan networks to discredit rivals, though individual clips’ hosting on YouTube versus other outlets varies by incident ^{[6] [4] [7]}.

5. When deepfakes touch state TV: the Putin martial‑law clip and the broader signal‑noise problem

A phony video of Vladimir Putin apparently declaring martial law in 2023 reached Russian television before authorities shut it down, demonstrating that synthetic media tied to major political events can leap from social platforms to broadcast ecosystems; not every influential clip was strictly native to YouTube, but the episode underscores the breach between online fakes and mainstream political discourse ^[8].

6. Scale, databases and the evidentiary gaps about platform provenance

Academic and NGO projects cataloged dozens to hundreds of political deepfakes from mid‑2023 through 2025 — for example, Insikt Group and the Political Deepfakes Incidents Database tracked scores of incidents across 38 countries — but many reports note inconsistent metadata, making it difficult to conclusively state that every cited clip was hosted on YouTube specifically even when YouTube is frequently mentioned among distribution channels ^{[4] [9]}.

7. Two important counterpoints: effects and “helpful” synthetic media

While alarmed analysts warned that deepfakes posed an escalating risk to elections, other researchers found a more nuanced picture: cataloguing 78 election‑related deepfakes, the Knight First Amendment Institute argued that not all deepfakes are malicious and that some served public‑interest or illustrative purposes, and rigorous causal links between specific clips and election outcomes remain sparse in the literature ^{[10] [7]}.

8. What the reporting supports and where uncertainty remains

Reporting from NGOs, think tanks and academic outlets confirms multiple high‑visibility synthetic videos and audio tied to elections between 2023–2025 and documents YouTube as one of the distribution platforms in several cases ^{[1] [2] [3]}; however, many sources acknowledge limitations in provenance data and sometimes conflate social‑platform spread with original hosting, so definitive, clip‑by‑clip attribution to YouTube is not uniformly available in the public record ^{[9] [4]}.