DF26 benchmark shows humans and top deepfake detectors are near random at spotting AI videos
ducha_aiki · x · 2026-09-10
- Researchers introduce DF26, a deepfake detection benchmark with 271 real and 2,420 synthetic videos of single-person public-speaking scenarios (direct-to-camera, official statements, studio interviews), all silent.
- Synthetic clips come from seven modern text-to-video / image-to-video models.
- Both humans and state-of-the-art deepfake detectors perform close to random chance — fakes are no longer reliably detectable.
- The authors argue current evaluation protocols are limited and call for benchmarks explicitly measuring robustness to modern generative model distribution shifts.
More from Safety
- How attackers get a shell without injecting code: chaining ROP gadgets — tetsuoai · 2026-09-10
- First-Ever Bill to Ban Superintelligent AI Introduced in UK Parliament, US Follows — GenericNameRandomNum · 2026-09-10
- BlueMoon exploit kit chaining Chrome and Windows zero-days spreads to 4 espionage groups — TechNadu · 2026-09-10
- Ex-DeepMind, now Anthropic researcher: no viable scientific plan for recursively self-improving AI risks — harris_edouard · 2026-09-10
- Zero CoT could let next-gen models scheme in latent space, security analyst warns — teortaxesTex · 2026-09-10
- Only Vendor Classifiers Hold Back Frontier AI Cyber Attacks — and Open Models Have None — AlexBarry4 · 2026-09-10