Deepfakes and AI-Generated Disinformation: A Threat to Democracy in the 2026 Elections

Abstract

Since 2024, the dissemination of electoral disinformation has been radically changed by the rapid emergence of affordable, easy-to-use generative artificial intelligence (AI) tools that can create believable-looking images, audio, and video of candidates and public figures saying or doing things that never actually happened. This paper draws upon recent academic research, government reports, and investigative reporting mostly from the past two years to explore the impact deepfakes and AI-generated disinformation pose to democratic processes as they enter the 2026 election cycle. The paper, based on case studies from the United States, Pakistan, India, Indonesia, Bangladesh, Canada, Ireland, Ecuador and the European Union, illustrates the extent, nature and impact of synthetic political content, particularly when used against women in public life. It explores the increasing difficulty of creating deepfakes versus combating their detection, the shifting nature of centralized platform fact checking to crowd-sourced approaches, the quick growth of state and international regulation, including the EU's Artificial Intelligence Act, and the psychological impact of deepfakes, which undermines trust in genuine evidence itself. The paper also looks at how trustworthy AI chatbots are as sources of election information and examines new election risks for the midterm and global elections of 2026. It ends with three levels of technical, regulatory and societal solutions and the view that resilience against AI-generated disinformation rests on a coordinated approach of increasing the capacity of disinformation detection, disclosure and media literacy among the public.