It's possible she doesn't keep up with technological developments within vocal synths, and agreed to record on the assumption that the end result would sound like a concatenative Vocaloid. Dr Yun is a poor communicator and rather unprofessional in a lot of his public-facing activities (ST Media has been in a fair amount of hot water with Korean fans over the years for how messily they're run), so I can envision a circumstance where she wasn't properly educated on what Uni would sound like, and didn't realise what the end result was until she stumbled onto it some time after Uni's release.
We do have a precedent for things like this happening, with Kafu also being blindsided by the realism of her SynthV and pulling out after she'd previously agreed to the point of demos going public. It would only have taken her company to not share demos with her, and for her not to have followed the development closely, for the same situation to have occurred.
From the point of view of hardcore fans, it's easy to forget that SynthV's output isn't common knowledge, and that remembering the different synthesis engines and methods is information 99% of people don't have. It's a reasonable possibility that the voice provider genuinely didn't predict, and subsequently isn't comfortable with, a perfect AI recreation of her voice.
I've lost count of the number of casual fans I've seen saying things like 'SynthV isn't AI', 'I thought Yi Xi was a real person!', 'I thought Teto was a Vtuber!', etc. It would only take a voice provider to a) Not realise what they're signing up to uses AI, or b) Not realise a demo sample they were sent was synthesised and not a recording example, to result in consent given under misinformed circumstances.
We do have a precedent for things like this happening, with Kafu also being blindsided by the realism of her SynthV and pulling out after she'd previously agreed to the point of demos going public. It would only have taken her company to not share demos with her, and for her not to have followed the development closely, for the same situation to have occurred.
From the point of view of hardcore fans, it's easy to forget that SynthV's output isn't common knowledge, and that remembering the different synthesis engines and methods is information 99% of people don't have. It's a reasonable possibility that the voice provider genuinely didn't predict, and subsequently isn't comfortable with, a perfect AI recreation of her voice.
I've lost count of the number of casual fans I've seen saying things like 'SynthV isn't AI', 'I thought Yi Xi was a real person!', 'I thought Teto was a Vtuber!', etc. It would only take a voice provider to a) Not realise what they're signing up to uses AI, or b) Not realise a demo sample they were sent was synthesised and not a recording example, to result in consent given under misinformed circumstances.