Explainable Deepfake Detection Challenge
TL;DR AI
2 min readKey summary
ACM Multimedia 2026 will host an explainable deepfake detection challenge built on XPlainVerse.
Participants must detect manipulated images and generate two explanations per case: a technical one and a simplified one for general users.
The benchmark scores both classification accuracy and the quality of grounded natural-language explanations.
The goal is to move deepfake detection beyond yes-or-no answers and toward trustworthy, evidence-based verification.