← All news

Safety

Survey update finds selected deepfake detectors struggle with unseen generators

An updated survey reports that, on its new out-of-distribution benchmark, the deepfake detectors evaluated by the authors often did not generalize to media from generators outside their training sources.

reviewedUpdated Aug 28, 2026, 2:20 AM UTC
Original source

arXiv

Read the original source

What happened

The authors updated “Deepfake Media Generation and Detection in the Generative AI Era: A Survey and Outlook,” with arXiv version 4 last revised on July 31, 2026. The paper surveys image, video, audio, and multimodal deepfakes.

The paper introduces BioDeepAV, an out-of-distribution benchmark, meaning it tests detectors on content outside their expected training sources. It includes more than 1,600 generated deepfake videos from four recent talking-face methods, plus audio-video examples with audio-only manipulation.

In their BioDeepAV evaluation, the authors report that selected state-of-the-art detectors failed to generalize to deepfakes from unseen generators. A separate AAAI conference paper found that frequency-based image detectors can overfit to artifacts in their training data and perform worse on unseen sources; it does not independently reproduce BioDeepAV’s audiovisual results.

Why it matters

The result is a testing implication, not evidence of failure in deployed authentication or safety systems. Cross-source testing can supplement a single accuracy score from familiar generators when assessing how a detector may perform against newly released generation tools.

What remains unclear

Sources

  1. Deepfake Media Generation and Detection in the Generative AI Era: A Survey and OutlookPrimary source - arXiv - research preprint / survey paper - Nov 29, 2024

    Used for: The updated survey, BioDeepAV benchmark description, and authors’ reported finding on unseen generators.

    Open source

  2. Frequency-Aware Deepfake Detection: Improving Generalizability through Frequency Space Domain LearningAssociation for the Advancement of Artificial Intelligence - peer-reviewed conference paper - Mar 24, 2024

    Used for: Independent, narrower support for the concern that frequency-based image detectors can overfit to known training sources.

    Open source

  3. Deepfake Detection that Generalizes Across BenchmarksarXiv - research preprint - Aug 8, 2025

    Used for: Counterpoint that stronger cross-benchmark generalization has been reported in tested settings.

    Open source