• Home
  • Blog
  • Android
  • Cars
  • Gadgets
  • Gaming
  • Internet
  • Mobile
  • Sci-Fi
Tech News, Magazine & Review WordPress Theme 2017
  • Home
  • Blog
  • Android
  • Cars
  • Gadgets
  • Gaming
  • Internet
  • Mobile
  • Sci-Fi
No Result
View All Result
  • Home
  • Blog
  • Android
  • Cars
  • Gadgets
  • Gaming
  • Internet
  • Mobile
  • Sci-Fi
No Result
View All Result
Blog - Creative Collaboration
No Result
View All Result
Home Internet

Hospitals adopt error-prone AI transcription tools despite warnings

October 28, 2024
Share on FacebookShare on Twitter

In one case from the study cited by AP, when a speaker described “two other girls and one lady,” Whisper added fictional text specifying that they “were Black.” In another, the audio said, “He, the boy, was going to, I’m not sure exactly, take the umbrella.” Whisper transcribed it to, “He took a big piece of a cross, a teeny, small piece … I’m sure he didn’t have a terror knife so he killed a number of people.”

An OpenAI spokesperson told the AP that the company appreciates the researchers’ findings and that it actively studies how to reduce fabrications and incorporates feedback in updates to the model.

Why Whisper confabulates

The key to Whisper’s unsuitability in high-risk domains comes from its propensity to sometimes confabulate, or plausibly make up, inaccurate outputs. The AP report says, “Researchers aren’t certain why Whisper and similar tools hallucinate,” but that isn’t true. We know exactly why Transformer-based AI models like Whisper behave this way.

Whisper is based on technology that is designed to predict the next most likely token (chunk of data) that should appear after a sequence of tokens provided by a user. In the case of ChatGPT, the input tokens come in the form of a text prompt. In the case of Whisper, the input is tokenized audio data.

The transcription output from Whisper is a prediction of what is most likely, not what is most accurate. Accuracy in Transformer-based outputs is typically proportional to the presence of relevant accurate data in the training dataset, but it is never guaranteed. If there is ever a case where there isn’t enough contextual information in its neural network for Whisper to make an accurate prediction about how to transcribe a particular segment of audio, the model will fall back on what it “knows” about the relationships between sounds and words it has learned from its training data.

Next Post

Apple Intelligence waitlist: How long it takes to get off the waitlist

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

No Result
View All Result

Recent Posts

  • Go West! US datacentres head for available and cheap energy
  • UK impact VC Eka Ventures closes second fund at $107M
  • I’m a Pixel fan, but the latest OneUI update is making me jealous
  • Nesto raises €11m from Expedition to scale AI workforce management for restaurant groups
  • 20x quota, long hours, or get out

Recent Comments

    No Result
    View All Result

    Categories

    • Android
    • Cars
    • Gadgets
    • Gaming
    • Internet
    • Mobile
    • Sci-Fi
    • Home
    • Shop
    • Privacy Policy
    • Terms and Conditions

    © CC Startup, Powered by Creative Collaboration. © 2020 Creative Collaboration, LLC. All Rights Reserved.

    No Result
    View All Result
    • Home
    • Blog
    • Android
    • Cars
    • Gadgets
    • Gaming
    • Internet
    • Mobile
    • Sci-Fi

    © CC Startup, Powered by Creative Collaboration. © 2020 Creative Collaboration, LLC. All Rights Reserved.

    Get more stuff like this
    in your inbox

    Subscribe to our mailing list and get interesting stuff and updates to your email inbox.

    Thank you for subscribing.

    Something went wrong.

    We respect your privacy and take protecting it seriously