Argentina’s 2023 presidential election, decided in a November runoff between economy minister Sergio Massa and libertarian Javier Milei, became an early case of AI image generation used heavily by both major campaigns. As documented in the OECD’s AI Incidents Monitor, the two sides used AI tools to produce deepfake videos and manipulated images attacking opponents, which spread widely on social media and contributed to a flood of election misinformation.
The imagery was often stylized rather than photorealistic. Massa’s team ran an Instagram account that recast him in heroic AI art, as a Roman emperor, a boxer, a film-style soldier, while portraying Milei and his allies as zombies and pirates. Milei, for his part, posted an AI image depicting Massa as an old-style communist in military dress that drew millions of views. More dangerous synthetic content also circulated, including a doctored video purporting to show Massa using drugs and a fake clip attributed to Milei that his rival’s campaign later disavowed. Milei went on to win the runoff with about 56 percent of the vote.
Argentina drew attention because much of the AI content was openly campaign-produced and even artistic, sitting in a gray zone between political cartoon and deception, rather than covert foreign interference.
For a business reader, it illustrates how fast generative imagery became a routine campaign tool, and how the mix of obvious satire and genuine fakes makes it harder for audiences to calibrate what to believe.