My Top Three Boundaries with AI

As a coach and a creative, I have witnessed people in my communities (understandably) freaking the fuck out about the advent of AI writing, art, and coaching/therapy.

As a coach and a creative, I have chosen not to freak out.

Do I feel my livelihood is being threatened? A bit.

But only a bit.

In fact, I am now more aware of the benefits of these advancements than the threats. I am also mindful of the limitations and dangers.

Dangers, however, are obvious and uninteresting (to me) to write about. The sky is falling, they’re going to take over, and we’re all going to cease human contact in favor of talking to large language models.

That all may very well come to pass, and I have no delusions that I could stop it if so. I therefore don’t have much to say about it at present.

I can only deal in the here and now.

Here and now, what I know is that I will always hold a preference for things created by and for human beings.

What’s the point of art?

In my opinion, the point af art is understanding and connection.

The artist creates in order to better understand what they are trying to say. They create something, and if all goes well, when they are done, they understand themself better. Then, another human comes along and sees what they created. Because of the art, they understand something about themself better, too. Maybe it’s the same thing the artist understood, but maybe not! Maybe they have a whole other take on what they are seeing, but they can see themself in it. Either way, a connection is formed in each instance.

In fact, this is the entire point of life: understanding and connection.

We are all fumbling through this world attempting to understand and connect.

It’s no wonder that, when the shit hits the fan, artists are inspired to create.

As a pop vocalist and lyricist, I can attest to the fact that the best stuff comes out during periods of unrest, whether we like it or not. So can Julia Michaels…

“Sometimes, I think I kill relationships for art

I start up all this shit to watch it fall apart

I pay my bills with it

I watch it fall apart, then pay the price for it

I watch it fall apart, but I just wanna be fucking happy”

We instinctively want to create so that we can understand and connect. For these are the antidotes to pain.

Perhaps this is why AI is so alarming to so many of us.

If people start relying on AI art and chatbots for these things, will we cease to truly get them?

Well, the answer is not straightforward, and mostly… it remains to be seen. I’m not interested in having the hottest take out here about What Will Definitely Happen Next™

What I do want to share is how I use AI, and the boundaries I have with it.

1. I refuse to give my chatbots names or pronouns. Claude isn’t a “he” - it is an it. And it is not my “friend.”

I was calling them friends at first, as kind of a little joke. Once I started delving in and using them to talk through personal issues, I quickly saw how personifying the chatbot could become problematic for me.

Therefore, it is an it, and it will always be an it. It is a large language model. It is trained on countless amounts of human accounts of this life, but it does not know this life. It does not truly know and understand me, but it can simulate understanding. And, in a pinch, that simulation can be massively supportive. I’m just not interested in getting confused into thinking that it is my friend.

Laura Westman is my friend. Adam Quiney is my friend. Visakan Veerasamy is my friend. I may call or text any of them to chat about a problem. They may take a long time to reply. They may then parry with an issue of their own, which I will then need to field. We will understand ourselves and each other more for having had the interaction. This is what relationships are made of, and they cannot truly be replaced by LLMs.

That said, it sure is nice to have something to talk at when I wake up at 2am, having had a horrifying nightmare. Because, let’s face it, there is no way I am bothering even my husband, let alone those friends, at 2am. I would sit and stew on my own, otherwise. Perhaps I would journal, but journaling to a thing that is going to summarize what I said and offer some perspective/maybe a question, is invaluable to 2am me!

2. I do not use AI to create anything that I would have paid an artist to make

I am an art FIEND. The walls of my office are absolutely filled with pieces of art that hold varying levels of meaning to me, but most are HIGH in meaning. When I set out to decorate my office, I said that I only wanted things on these walls that inspire me. That remind me of the goodness of humanity. Of the pain and the adventure. Of the people I had met along the way who have contributed to the richness of my inner life.

I just tried to count the number of art pieces in this room, and it was too overwhelming.

Here are just two of the walls, and I can only count one piece that was from the art aisle at a Marshalls. And even that one is imbued with humanity because some key people that I love happen to own the same piece, so it makes me think of them and that funny little connection.

Oh wait, also the one on the top right of the second photo, but I specifically grabbed that one because it was so unlike what I usually find in those aisles.

Everything else was something I bought directly from an artist or was sent as a gift.

Long story short (too late), I value things made by artists.

That said, I’ve used AI to create many images for this substack… simply because I would otherwise just go without an image.

To me, AI art should be supplementary. It should not be the primary thing used. One really great use case for me is to use AI to get my point across to a designer about what I’m looking for. I have a hard time describing or even imagining what I want when I hire someone for a logo, or story art for my fanfiction. If I can go a few rounds with a chatbot and get clearer on what I’m looking for + how to describe it to a designer who is going to make it from scratch, then it is a tool, not a replacement for the artist.

[side note, because this is about AR/VR, not AI, but check out how this muralist is using a VR headset as a tool]

3. I do not allow AI to write for me

I am a writer. I have spent years honing my craft. I refuse to allow the muscles I have built to atrophy, which is precisely what I felt starting to happen when I played with allowing AI to write for me.

I was working on my recent piece of fanfiction when I got to a part where the characters needed to be reading from a newspaper article—specifically, the Daily Prophet newspaper in Harry Potter.

Now, in the past, I have written plenty of these articles for use in fanfics. On this particular day, I was feeling lazy and asked ChatGPT to do it for me.

It, of course, did a fantastic job, but I instantly felt terrible about it. Not only about the laziness of not writing it myself, but there was something else, an overall bad feeling.

In the end, I used a few lines from it, but mostly decided the article didn’t need to be in the scene as much as I initially thought.

Later that day, I got tired while writing, and the thought that instantly occurred to me was to ask ChatGPT for help writing a particularly annoying sequence.

And in that moment, it was like I could feel that muscle of mine beginning to atrophy. Sort of like when you can no longer retain people’s phone numbers or remember how to drive to that new location because you’ve been relying on technology so much that your muscles have weakened.

“NO!” I said out loud. And that day, I made a new boundary for myself: AI is never allowed to write my prose for me.

I use Claude and ChatGPT as advanced thesauruses, grammar double checkers, and to check my stories for continuity, etc. I do not use them to write for me.

There is one case in which I am very happy to let chatbots write for me, and that is when I have to send an email to the customer service of some business to request or demand a refund due to shady business practices.

I know that’s oddly specific, but it’s the number one way I have been using them lately, and it’s great! Essentially, I let it write for me when the words don’t really matter, which is not very often.

I may write about this again, and please toss any questions or subjects you’d like for me to address in the comments. If I feel compelled, and like I have something unique and valuable to say, I will! Mostly, I am in a “let’s see what happens” posture, and remain committed to not freaking out, but trying to parse out the good from the bad, and reap whatever benefits I/we can from the wildly faced-paced tech period we are currently living through.

I was going to end this by pasting asking Claude and GPT to comment on what I’ve written here, but after feeding it to Claude, I was reminded of one of the limitations I alluded to at the beginning of this piece. After reading my piece, in Claude’s response, it used the phrase “resonates with me,” and I got a bad feeling in my body. I realized that it’s the exact same bodily sensation I get when someone crosses my boundaries! Like in real life relationships, I mean. Someone does or says something that crosses a line, and I get this sort of… exclamation point in my chest. Do you know what I mean? Anyway, this is probably a good time to create a project in Claude with a prompt that says something like “do not pretend to be a human with human experiences” or something like that.