I'm a lawyer, right? That means I actually get legal work done. And a lawyer who isn't using AI at this point is like a woodcutter not using a chain saw. Makes no sense. However, a careless or inexperienced lawyer using AI can easily carve off a limb, and not realize their error. So I've been working with ChatGPT and Gemini for a couple of years, during which time I've written a couple of books I haven't published, and a fair stack of legal work using Lexis AI, that they call Protege.
Writing fiction with ChatGPT was interesting, but left to itself, it clearly had no creative capacity at all. If I wanted a rich scene description, I had to puff a lot of imagery and atmosphere into the prompt.
AI is much better at the routine digestion of information from documents, chronologizing facts into a narrative, and applying an analytical template supplied by the prompt. Research memoranda generated in response to well-formed prompt are usually reliable, but if you don't know the field of law well, you can easily be taken in by a plausible, wrong AI conclusion.
The AI-using lawyer must diligently sift the AI slop, confirming factual narratives and chronology, checking legal citations to statutes, regulations, and cases, squeezing out the facile assumptions and insisting on getting specific answers to questions about dates, times, identities, money movements, etcetera. For me, it's a lot like having a pretty talented, but rather inexperienced associate who usually produces good work, but must be carefully instructed, and his work diligently checked, to be useful. The great danger of course is that both the lawyer and their AI assistant will be consumed by a Rumsfeldian "unknown unknown." Then you end up having your adversary notify you of your error, which is of course a very bad thing.
So that's how I use AI professionally. How do I use it on this website, and why? I share the results of my exchanges with AI chatbots, often without any editing, because I like getting answers to my questions, and AIs are pretty amazingly good at providing correct answers. And since I can dig into any topic in greater detail using further AI prompts and standard search techniques, I feel it serves me very well when I'm trying to get the skinny on issues and events that the mainstream media don't focus on, or even conceal from public knowledge.
I always identify AI-created work that is posted unrevised. I often write things from scratch after reading AI-generated research. That' my work of authorship, and my use of AI research doesn't alter its originality. Sometimes I tweak the AI's output to make it more readable, but I'll still identify it as AI slop. But if it gets to the point of a virtually total rewrite, I credit myself.