paper
Ethical and Technical Limits of Deepfake Speech Datasets
paperactiveprovisional
ethical-and-technical-limits-of-deepfake-speech-datasets-a8752504·1 events·first seen 7d agoAliases: Ethical and Technical Limits of Deepfake Speech Datasets
More like this (12)
What Do Deepfake Speech Detectors Actually Hear?Leveraging Audio-LLMs to Filter Speech-to-Speech Training DataGiving AI a Headache: Acoustic Adversarial Attacks to Computer Vision ApplicationsFrom Self-Supervised Speech Models to Mixture-of-Experts for Robust Anti-SpoofingOpenAI voice modelsOpenAI Voice AICan News Predict the Market? Limits of Zero-Shot Financial NLP and the Role of Explainable AITrustworthy AIelection deepfakesEU Code of Practice on AI content transparencyAdversarial Creation and Detection of AI-Generated Social Bot ContentBeyond task performance: Decoding bioacoustic embeddings with speech features
Recent events (1)
Audit of 39 deepfake speech datasets reveals fairness and generalization gaps
A dataset-level audit of 39 deepfake speech datasets examines accessibility, documentation, demographic coverage, scale, and source corpora. The study finds that fairness assessment is largely infeasible due to missing demographic metadata, and that substantial overlap in underlying speech corpora across datasets undermines cross-dataset evaluation and inflates generalization claims. The findings challenge the credibility of robustness and fairness claims made for deepfake speech detectors.