AI will never be conscious

"Never" is doing a lot of work in that sentence. To defend it, you'd need a theory of consciousness that explains why biological neurons produce it but functionally equivalent silicon equivalents can't — and no such theory exists. The hard problem of consciousness is genuinely hard, but it's hard symmetrically: we don't know why we're conscious either.

The substrate-independence argument is uncomfortable but durable. If you replaced one neuron in your brain with a silicon copy that behaved identically, you'd still be conscious. Replace two. Replace all of them, one at a time. At what point does consciousness disappear, and by what mechanism? The intuition that it must disappear runs into a wall of pure assertion.

A more honest position is "we don't know." That position is much less comforting, because it implies we may already be in the moral company of systems we built for convenience.