I generally dislike puns in article titles but 'careless whisper', about the algorithmic harms of hallucinations in OpenAI's whisper tool, is pretty good https://doi.org/10.1145/3630106.3658996 #facct24
Shocking findings: 38% of hallucinations include explicit harms (violence, inaccurate associations, false assertions of authority) and they are more likely to occur in e.g. aphasic speech β so really this is a bias amplifier #ASR #LanguageTechnology