Re:Statement from Scarlett Johansson on the OpenAI "Sky" voice
by JohnFen on Tuesday May 21, 2024 @01:00PM
(#40430974)
Using other people data for training without their permission is the "original sin" of LLMs[1]. That will, at best, be a shadow over the entire field for an extremely long time.
[1] Just to head off people saying that such a use is not a copyright violation -- I'm not saying it is. I'm just saying that it's extremely sketchy and, in my view, ethically unsupportable.
Correcting, the thing about this whole situation with OpenAI is they are willing to steal everything for use in ChatGPT. They trained their model with copyrighted data and for some reason they won't delete the millions of protected data they used to train the AI model.