How deepfakes undermine truth and threaten democracy

A Deepfake Video Disrupts an Election Night

Picture a tight race on election eve where a video surfaces showing a leading candidate confessing to corruption. Within hours the clip races across social platforms, voters share it in group chats, and poll numbers shift before fact-checkers can respond. The footage looks authentic, sounds convincing, and leaves citizens unsure what to believe as they head to the polls.

Danielle Citron's Central Argument

Law professor Danielle Citron's talk, "How deepfakes undermine truth and threaten democracy," centers on the claim that AI-generated deepfakes spread misinformation, manipulate public perception, and disrupt democratic elections by eroding trust in truth. She argues that these synthetic media do more than deceive individuals; they attack the shared foundation citizens need to weigh candidates and policies. By making people doubt the very possibility of reliable evidence, deepfakes weaken the collective ability to hold elections accountable.

Citron develops this claim by tracing how deepfakes operate at scale. Once created, they travel faster than corrections and leave lasting doubt even after they are exposed. The result is not simply confusion about one video but a broader skepticism toward all recorded statements, including genuine ones. In her view, this erosion of epistemic trust strikes directly at democratic processes that depend on voters' capacity to discern fact from fiction.

How the Argument Illuminates the Scenario

Returning to the election-night video, Citron's analysis shows why the damage extends beyond any single falsehood. The clip does not merely change a few minds; it plants the suspicion that every piece of evidence could be fabricated. Voters who later learn the video was fake may still question authentic footage of the same candidate or of opposing voices. The speaker's emphasis on eroded trust explains why simple debunking often fails to restore confidence once deepfakes have circulated. It also underscores the urgency of safeguards that address detection and mitigation before such material reaches voters, precisely the need highlighted in her discussion of protecting electoral integrity.

A Question That Lingers

If deepfakes can make people believe something is real when it isn't, striking at the heart of our ability to discern truth from fiction, what coordinated steps by governments and technology companies can restore enough shared certainty for elections to remain credible?