In October, a fabricated clip of Keir Starmer swearing at staff was shared. Imagine the impact if it had happened just before an election
In March 2022, a video emerged of Ukrainian President Volodymyr Zelensky, appearing pale and weary, telling his soldiers to lay down their weapons and surrender. The clip, however, was a deepfake—an AI-generated video designed to deceive viewers into believing the real Zelensky was capitulating to Russia’s invasion. Although quickly taken down by social media platforms, the incident raised a chilling possibility: What if the video had been more sophisticated or timed during a communication blackout in Kyiv? It could have potentially altered the course of the war.
This was not an isolated case. In the run-up to Turkey’s 2023 presidential election, President Recep Tayyip Erdoğan shared a fabricated video showing his rival, Kemal Kılıçdaroğlu, receiving an endorsement from the Kurdistan Workers’ Party (a designated terrorist group in Turkey). While Kılıçdaroğlu debunked the video, it had already spread widely, influencing voters’ perceptions before the election. Erdoğan won the subsequent run-off, securing his third decade in power.
Similarly, in the United States, political deepfakes have become increasingly common. For example, in June, Florida Governor and presidential candidate Ron DeSantis’ team released a campaign ad showing Donald Trump hugging and kissing Dr. Anthony Fauci, the former chief medical adviser to the White House. The video mixed real footage with AI-generated elements, presenting a misleading and inflammatory message. Joe Biden has also been targeted, with a deepfake video circulating in February, showing him announcing a national draft for Americans to fight in Ukraine, further complicating political discourse.
As AI-generated content continues to improve, the potential for deepfakes to disrupt democratic processes grows. Fabricated videos can be spread quickly, making it difficult for political leaders to address misinformation before it has done significant damage. As demonstrated in these cases, deepfakes have the potential to create widespread confusion, manipulate public opinion, and undermine trust in democratic institutions, especially when such content is released just before critical elections. The question remains: In an era of deepfakes, how can democracy be protected?