Tech behemoth OpenAI has touted its artificial intelligence-powered transcription tool Whisper as having near 'human level robustness and accuracy.' But Whisper has a major flaw: it is prone to making up chunks of text or even entire sentences.
The experts said some of the invented text - known in the industry as hallucinations - can include racial commentary, violent rhetoric and even imagined medical treatments, revealing significant concerns over the tool's reliability.
Experts expressed concern about the use of Whisper in medical contexts, noting that institutions are rushing to adopt Whisper-based tools to transcribe patient consultations despite clear warnings against using it in 'high-risk domains.'
A researcher found hallucinations in eight out of ten audio transcriptions of public meetings, while a machine learning engineer reported finding hallucinations in roughly half of the 100 hours of Whisper transcriptions analyzed.
Collection
[
|
...
]