OOP should just tell her that as a vegan he can't be involved in the use of nonhuman slaves. Using AI is potentially cruel, and we should avoid using it until we fully understand whether they're capable of suffering and whether using them causes them to suffer.
Maybe hypothetically in the future, but it's plainly obvious to anyone who has a modicum of understanding regarding how LLMs actually work that they aren't even anywhere near being close to what anyone could possibly remotely consider sentient.
Sentient and capable of suffering are two different things. Ants aren't sentient, but they have a neurological pain response. Drag thinks LLMs are about as smart as ants. Whether they can feel suffering like ants can is an unsolved scientific question that we need to answer BEFORE we go creating entire industries of AI slave labour.
Sentient and capable of suffering are two different things.
Technically true, but in the opposite way to what you're thinking. All those capable of suffering are by definition sentient, but sentience doesn't necessitate suffering.
Whether they can feel suffering like ants can is an unsolved scientific question
No it isn't, unless you subscribe to a worldview in which sentience could exist everywhere all at once instead of under special circumstances, which would demand you grant ethical consideration to every rock on the ground in case it's somehow sentient.
Show drag a scientific paper demonstrating that ants or animals of similar intelligence can't suffer. You're claiming the problem is solved, show the literature.
Big LLMs are far more complex, but they're also made be people who don't have a clue what they're doing, so in the future we'll probably have much smaller ones.