ASVspoof 5: Evaluation of Spoofing, Deepfake, and Adversarial Attack Detection Using Crowdsourced Speech

2026-01-07

AI Summary

This study addresses the significant lack of robustness in current speech anti-spoofing systems under complex scenarios such as deepfakes, adversarial attacks, and neural codec compression. To this end, the work introduces the first large-scale crowdsourced speech database encompassing diverse speakers, recording conditions, and both legacy and state-of-the-art speech synthesis techniques. The authors systematically evaluate the generalization capabilities of 53 detection methods across these challenging conditions. Experimental results demonstrate that while most approaches perform well under standard settings, their performance degrades substantially under adversarial attacks and neural compression. The study further uncovers critical bottlenecks—such as model calibration—that hinder robustness, thereby establishing a foundational dataset and a clear research roadmap for advancing speech anti-spoofing technologies.

Technology Category

• Machine Learning: Adversarial Learning & Robustness
• Natural Language Processing: Speech
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Security and Privacy: Large-scale security measurements
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

ASVspoof 5 is the fifth edition in a series of challenges which promote the study of speech spoofing and deepfake detection solutions. A significant change from previous challenge editions is a new crowdsourced database collected from a substantially greater number of speakers under diverse recording conditions, and a mix of cutting-edge and legacy generative speech technology. With the new database described elsewhere, we provide in this paper an overview of the ASVspoof 5 challenge results for the submissions of 53 participating teams. While many solutions perform well, performance degrades under adversarial attacks and the application of neural encoding/compression schemes. Together with a review of post-challenge results, we also report a study of calibration in addition to other principal challenges and outline a road-map for the future of ASVspoof.

Problem

• speech spoofing
• deepfake
• adversarial attack
• crowdsourced speech
• detection

Innovation

• crowdsourced speech
• deepfake detection
• adversarial attack
• speech spoofing
• neural encoding