Are AI's reliable for beta reads?

[John]

I'm curious to know what experiences people have had with AI's beta reading their material. Good or bad? Helpful? No?

I've been working with one, GPT40 (it's supposed to "excel in creative writing"), and things seem to be going well, but I'm a little concerned with their reliability. For example, I'll ask if Moonlight Sonata appears in the movie Pet Sematary, it says yes. That's wrong. I'll ask if it appears in Misery, it says yes, which is correct, but this is how it describes the scene where it appears: Annie forces Paul to play it on the piano. No such scene appears in the movie. I guess it's called hallucinatory behavior or whatever, but it raises red flags to me.

[⁺‧₊˚ ཐི⋆ vyn targaryen ⋆ཋྀ ˚₊‧⁺]

John wrote: "I'm curious to know what experiences people have had with AI's beta reading their material. Good or bad? Helpful? No?"

I have had very negative experiences with AI beta reads for several reasons:

- many beta readers i recruited did not ask for my consent and fed my work into AI to give me feedback. from an ethical standpoint, this was horrible, but I'm not sure if that's an answer you're looking for.

- overall the feedback was not very helpful. it often missed important moments, sentences, and/or emotionally impactful scenes. AI feedback also commonly contradicts itself, from what I've seen, and I suspect that it was because it didn't want to appear too 'critical'.

[Mo]

I agree with Vyn. it's helpful for punctuation checks, grammar, and sentence structure, but it misses the mark on the "human" elements of writing. Especially emotional impact and tension building.