Deepfakes in Political Manipulation: Evaluating Risks Under the AI Act

Deepfake technology poses a major threat to democratic processes, particularly elections. In the USA, deepfakes were expected to influence the 2024 presidential election, while in India they were used during the 2024 general elections to sway voters. European countries like Poland and Slovakia also experienced deepfake-driven campaigns undermining political narratives. These cases highlight how low-cost AI tools can industrialize disinformation at an unprecedented scale. This paper examines political deepfakes under the AI Act’s systemic risk framework, evaluating intent, scale, and impact, and proposes AI detection systems and regulatory measures to enhance transparency and fairness. By combining case studies with regulatory theory, it emphasizes the urgent need for legal, technological, and social countermeasures.