> "And offloading that labor, regardless of whether it's taxing or not, is undercutting my ability to be a moral actor."

The objection in the article is specifically about offloading thinking to technology, which is something we've been doing for decades, or depending on your definition, since we first started using tools.

If offloading cognitive labor is the core problem, then what makes AI fundamentally different from the countless other technological advances that have allowed us to offload cognitive labor?

Exactly. Recent LLMs are in a different category and it's disingenuous to pretend they're not. Wherever you decide to draw the line, GPT-3 and more advanced chatbots are clearly on the far side of it.

Sorites doesn't actually refute the OP. The paradox is rooted in a failure to distinguish between aggregates and integral wholes. A heap is not really a natural kind or an integral whole. It is just a convenient label to describe a vague range of collections for practical purposes, but it is not what you a "thing" in its own right. There is no intrinsic tendency of a heap to be a heap, whereas in integral wholes there are. You cannot draw a line between a heap and a non-heap, because there is no such thing as a heap in any objective sense. There is no essential difference, only an accidental one.

Computers and computer programs are much more like heaps than integral wholes. There is no essential difference between what we conventionally call "AI" and any other computer program. The manner in which we have cordoned off "AI" from "non-AI" has more to do with human purposes and perspectives rather than any objective facts on the ground.