A report by the Oxford Internet Institute found that AI-generated deepfake content has been deployed in 78% of national elections held worldwide in 2026, up from 34% in 2024.
The deepfakes range from fabricated speeches by political candidates to manipulated images and audio designed to spread misinformation. Detection technology has improved, but distribution through encrypted messaging apps makes containment nearly impossible.
The most affected regions are South Asia, Southeast Asia, and Sub-Saharan Africa, where digital literacy is lower and fact-checking infrastructure is limited.
Social media platforms have deployed AI-powered detection systems, but the arms race between generation and detection favours the attackers. Detection accuracy rates hover around 85% — meaning 15% of deepfakes evade automated screening.