Media Provenance as a Prevention for Malicious Synthetic Media
summary
Malicious synthetic media includes both deepfakes and cheapfakes; creation tools move from a $20,000 Photoshop-era workflow to $24-per-month subscriptions, making fakery widespread. Detection exists but is an arms race and achieves only 60–75% detection accuracy, so false positives and negatives worsen the Liar’s dividend. Provenance attaches a manifest with four parts (asset fingerprint, metadata, edits, signature) cryptographically bound by hashing and signatures, validates quickly even for live streams, and is being standardized via C2PA standard to provide verifiable origin signals.
AI-written summary · about
chapters (10)
- 0:00Media provenance basics
- 1:51Synthetic media risks
- 3:59Malicious media examples
- 7:11Why detection alone fails to solve the problem
- 10:19Provenance versus detection
- 13:30Provenance manifests
- 16:41Provenance standards leaders
- 20:58C2PA future
- 21:41Key takeaways
- 22:02Q&A
related talks
41mThe Deepfake Supply Chain: Weaponizing Synthetic Media into Modern Cybercrime Operation
30mAI-Generated Synthetic Media, aka Deepfakes and its impact
12mUnmasking Digital Deception: The Content Authenticity Initiative For Video Engineers
27mTruth or Fiction The Art of Unveiling DeepFakes in Text, Audio, Image
14mContent Credentials - Identities for Media Content
27mIs now the time to solve the deepfake threat?
30mReality Under Attack: Just How Serious Is the Deep Fake Threat?
26mDeepfakes, audiofakes, and the future of media