Oct 20, 2026 · 30 min
Deepfakes threaten evidence by making truth easy to dismiss
If deepfakes are getting better, can we trust our lying eyes?
As synthetic media improves, the central problem shifts from spotting fakes to preserving trust in authentic evidence before misinformation spreads.
- 1Deepfake detection struggles against the speed and scale of online distribution, even when forensic analysis eventually establishes the truth.
- 2Authentic recordings can now be dismissed as fabricated, turning uncertainty about synthetic media into a defense against accountability.
- 3Human perception performs poorly on emotionally charged AI-generated media, making platform amplification and institutional verification increasingly important.
Don't miss
Farid explains how analysis of trajectories, distances, and the delay between a flash and explosion supported the authenticity of shaky missile footage.
The brief
Digital forensics expert Hany Farid says deepfakes are becoming harder to detect, but the deeper threat is that authentic evidence can now be dismissed as fake.
Farid recounts investigating shaky footage that appeared to show a Tomahawk missile approaching a girls’ school in Iran; trajectories and timing supported its authenticity.
The internet can spread a false interpretation to millions before investigators establish what happened, while disproving manipulation is harder than simply finding no evidence of it.
Voice and face cloning now require little source material, and emotionally charged media makes people especially poor at recognizing fabricated evidence.
Farid recommends family code words for suspicious calls and professionally edited news sources, while arguing that platform liability may be needed to curb amplification.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.