In general, AI text should not be used verbatim. In writing or in speech. Think of it as lorem ipsum. Or more precisely, lorem aipsum. It's placeholder text.

All this may change as tech evolves. Perhaps everyone gets their own highly personalized writer, such that the AI copy they emit isn't cliche-ridden and overwritten. But at least for now...there's an instant turnoff when you see someone use AI text verbatim in an application form, or a Substack post, or a tweet. It just signals that they can't write well, or that they're lazy, or something. To be clear, AI is useful for coding, for data analysis, for snippets of legal work, for situations where there's a right answer. It's useful for searching the web, for suggesting synonyms, for cleaning up audio transcripts, for summarizing huge docs, and in general as an assistant. But it's not useful as a replacement for your writing. As soon as you read a few sentences of AI text, it just clangs on the ear. I have to say that I kind of get why some artists hate AI. I think AI images/video are extremely useful for someone who can't draw to get something out of their mind's eye onto the page, but I get why artists who identify themselves with their art dislike it. AI art makes it harder for them to stand out. Similarly, AI-generated text is probably very useful for someone who speaks English as a second language (or just isn't that good as a writer). It helps them put together an essay from their bullet points. I get all that, and there are situations where it's very reasonable to use (eg realtime translation). But unless the AI text is better than your best writing, you really shouldn't use it verbatim in high-stakes contexts.