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, according to interviews with more than a dozen software engineers, developers and academic researchers. Those experts said some of the invented text — known in the industry as hallucinations — can include racial commentary, violent rhetoric and even imagined medical treatments.
Experts said that such fabrications are problematic because Whisper is being used in a slew of industries worldwide to translate and transcribe interviews, generate text in popular consumer technologies and create subtitles for videos.
More concerning, they said, is a rush by medical centers to utilize Whisper-based tools to transcribe patients’ consultations with doctors, despite OpenAI’ s warnings that the tool should not be used in “high-risk domains.”
maybe that’s just the ai’s internal monologue leaking through?
Why is generative AI even needed for audio transcription? We’ve had decent voice recognition tools for years even on cheap consumer grade stuff.
We also used to have decent web searches!
Because with normal algorithms you have someone to blame.
AI is a trick to hide when you steer the results the way you want.
Whisper really is a lot better when it works, and it’s free. The problem is that it refuses to produce gibberish or give up when it doesn’t work. You’ll always need an editor.
Some examples
In this example, the speaker said, “as the um, the, her father dies not too long after he remarried….” while the program transcribes that as " It’s fine. It’s just too sensitive to tell. She does die at 65….”
In this example, the speaker said, “and after she got the telephone he began to pray” while the program transcribes that as “I feel like I’m going to fall. I feel like I’m going to fall, I feel like I’m going to fall….”
Wow, that’s bad. I thought it would be more of a “confusing a sentence for a similar sounding one” type thing but from the above and the article it’s just generating semi-believable text and sticking them into the transcriptions.
It’s actually extremely good at figuring out confusing text. It gets weird when the audio quality is bad.
I use it for generating subs for obscure movies.
LLMs in medicine. What could go wrong?
“This seems solvable if the company is willing to prioritize it.”
I know how to make the company prioritize it: make it illegal to use (or even promote) until a certain threshold of accuracy is met. This software is absolute garbage at best, and a genuine hazard at worst.
Lame, ineffective “warnings” serve no purpose but to cover OpenAIs ass. Hit them in the wallet, and they’ll pay attention.
Microsoft teams has some automatic transcript capabilities that are so hilariously bad, it’s hard to believe Microsoft released it.
I guess they use the same service.
I’ve read that this is only going to continue to happen (and get worse) because we’re basically out of human-generated training data that’s publicly available on the internet, so models are being trained on content generated by other models. They literally make shit up constantly, and every generation gets dumber and dumber until they can’t even stay on topic or complete a coherent sentence anymore.
Odd that GPT (and of course all the LLMs too) only got better so far…
The architecture changed, there is still progress to be made there. But LLMs will forever be stuck in 2021, all data afterwards is tainted. Not a lot has been added.
In fact, Whisper was developed to transcribe videos for more training data, because they ran out of text data. These bad transcriptions are in newer models.
AI does not mean Artificial Intelligence, it means Assumed Intelligence.
Oh word?
