Comment by ingonealan3

Your line of thinking rests on the assumption that "AI enables machines to understand English", but this is false. AI, or more precisely LLMs, are a class of algorithms that take a string as input, and return a new string generated through statistical black magic. They don't enable the machine to do anything in particular. You can't ask the LLM to run some arbitrary computation directly on the CPU, for example. If that were possible, your argument would hold some water, at least to some extent.