I had a comment said to me the other day that made me ask, "do people think I use AI to write?". Now this comment wasn't directly insinuating that I do, but more of directing me not to write with AI for an idea.
To give the context, I had presented an idea to my senior pastor about making some training materials for our campus pastors to be equipped for managing church finances. It's one thing to survive or even thrive in personal budget and expenses, but it is another to manage the budget for a church. When you're fully reliant on people's generosity, you don't have your own venue so you are renting essentially on a week to week basis, you have a number of expenses that you aren't really in control of like insurances, AND you are most likely bivocational so this isn't even your full time job but you need to think ahead and budget and estimate future costs or expenses - there's a lot to learn and look after.
My pastor liked the idea and gave me the go ahead to present to him a training outline - topics and contents that I think would be beneficial for everyone to be taught from. But he gave me this instruction: "no 'this stupendous and fantastic yada yada yada' kind of language. Don't send me something made with AI, my brain can't handle it, I just need short and clear." Reality of that comment is that he's probably sick of seeing plans or content sent to him that directly came from a prompt to an LLM asking it to "make me a plan for blah blah blah" with a message that says something like "I had this idea, this is what Chat gave me". But, that little niggle of a question around what people think of my writing was enough to make me ponder.
What are some reasons people would think I use AI to write? Well for starters I do love my em dashes - and apparently LLMs do too and so it's a bit of a "tell" for some people. I also do like to explain myself and be clear with reasoning and LLMs create their responses based on reasoned learning, even if it is sometimes hallucinated. Now personally I actually hate how "fluffy" an LLM's explanation can get, just give it to me straight man. Sound like a real person. It's a critique I have with some people trying to overcompensate hyping something up instead of knowing the foundational why and motivation people need to be moved to act on something. It's also the thing I most commonly do correct an LLM on when I use them for translation. I so often have to tell it "don't add to my message, just translate what I said" even though that's part of the clear instruction I have for the chat.
The last reason I was prompted to ask myself that question, is I have seen other people who often write get accused by people unfamiliar with them of using AI to write. I read their tweet or their blog post and because I am familiar with them and their own disdain for AI writing wouldn't think for a second it was AI slop. But they get these ignorant people commenting and accusing them of just copying and pasting from Chat. If these people, who regularly write, who have been very vocal about hating AI slop are getting accused of posting it...well maybe it's not that far fetched that the people I regularly write to or for would mistake some of what I write as straight from an LLM too.
Am I the best writer in the world, definitely not. But I take solace in this: if AIs have been trained on the best written material on how to sound like a thoughtful human, and some of how they write is mirroring how I have written since before anyone would ever think G and P and T are letters that go together, maybe my written communication is pretty good and human-like after all.
To give the context, I had presented an idea to my senior pastor about making some training materials for our campus pastors to be equipped for managing church finances. It's one thing to survive or even thrive in personal budget and expenses, but it is another to manage the budget for a church. When you're fully reliant on people's generosity, you don't have your own venue so you are renting essentially on a week to week basis, you have a number of expenses that you aren't really in control of like insurances, AND you are most likely bivocational so this isn't even your full time job but you need to think ahead and budget and estimate future costs or expenses - there's a lot to learn and look after.
My pastor liked the idea and gave me the go ahead to present to him a training outline - topics and contents that I think would be beneficial for everyone to be taught from. But he gave me this instruction: "no 'this stupendous and fantastic yada yada yada' kind of language. Don't send me something made with AI, my brain can't handle it, I just need short and clear." Reality of that comment is that he's probably sick of seeing plans or content sent to him that directly came from a prompt to an LLM asking it to "make me a plan for blah blah blah" with a message that says something like "I had this idea, this is what Chat gave me". But, that little niggle of a question around what people think of my writing was enough to make me ponder.
What are some reasons people would think I use AI to write? Well for starters I do love my em dashes - and apparently LLMs do too and so it's a bit of a "tell" for some people. I also do like to explain myself and be clear with reasoning and LLMs create their responses based on reasoned learning, even if it is sometimes hallucinated. Now personally I actually hate how "fluffy" an LLM's explanation can get, just give it to me straight man. Sound like a real person. It's a critique I have with some people trying to overcompensate hyping something up instead of knowing the foundational why and motivation people need to be moved to act on something. It's also the thing I most commonly do correct an LLM on when I use them for translation. I so often have to tell it "don't add to my message, just translate what I said" even though that's part of the clear instruction I have for the chat.
The last reason I was prompted to ask myself that question, is I have seen other people who often write get accused by people unfamiliar with them of using AI to write. I read their tweet or their blog post and because I am familiar with them and their own disdain for AI writing wouldn't think for a second it was AI slop. But they get these ignorant people commenting and accusing them of just copying and pasting from Chat. If these people, who regularly write, who have been very vocal about hating AI slop are getting accused of posting it...well maybe it's not that far fetched that the people I regularly write to or for would mistake some of what I write as straight from an LLM too.
Am I the best writer in the world, definitely not. But I take solace in this: if AIs have been trained on the best written material on how to sound like a thoughtful human, and some of how they write is mirroring how I have written since before anyone would ever think G and P and T are letters that go together, maybe my written communication is pretty good and human-like after all.