As a writer, I am a bit offended by AI.
I suspect that artists and developers probably struggle with similar issues – what AI creates, is OK, but it’s not great, and any specialist in a certain field can still spot its cold, metallic fingerprints.
Nowhere is this more evident than on social media
Here is how I read the average AI generated slightly-salesy, highly-pandering post:
I had AI create me a post about posts, and I realized something incredible:
It didn’t just write a short sentence.
It didn’t even write another short sentence.
It wrote a longer sentence that completely changes the way sentences are written, from the bottom up. Period.
The post it created doesn’t just explain (italics for emphasis) any old idea. It it didn’t just add substance to an existing (italics for emphasis) idea. It’s a hyperbolic new way to state an idea that has never been done before (3rd set of italics really drives the point home).
That’s the brilliance of this short sentence that emphasizes previous one.
The genius of AI is that the final closing line really ties it all together all the sentences combined in a way that says “wow, I definitely couldn’t have written than better myself”.
And for that, I applaud you.
I’ve been wanting to get that off my chest off a long time.
Thanks for bearing with me.
AI Shortcomings
I criticize AI even as I use AI everyday to ship new tech, which has actively been used in driving revenue for businesses.
AI is very good at some things.
It’s great at spotting patterns.
It’s great at organizing thoughts.
It’s great for creating something good-enough, and good enough is often all we need.
For now, at least, AI’s results are mediocre at best, and any professional is able to spot this. The takeaway, somewhat sadly, is that mediocrity is often sufficient for a variety of situations.
A mediocre social media post is enough to get engagement.
A mediocre image is enough to convey a point.
As an artist I’ve always struggled with this facet of the human condition.
I think embracing it, understanding when good-enough is ok and when to keep working to make something perfect, is part of maturity and the inherent pain of living in an imperfect world.
The Humanity of Communication
Maybe I’ll still split an arbitrary hair just to make myself feel more self-important. I would argue that personal communication should not involve AI.
Communication should be the sharing of self, and running it through a filter is a dilution of authenticity.
It is not that different than posting a photo of yourself through a filter. I don’t want to see a filtered version of you, I want to see you.
Some communication is technical, and all it needs to be is dry and clear.
But a lot of communication is about shared humanity, which social media has already reduced to virtual, rewards-based systems.
Nonetheless, a well-written post or article can still inspire, move, or touch us. AI threatens to further dehumanize whatever remains.
Technically communicating product specifications can involve AI. Expressing how you feel about a frustrating situation, should not.
To me, there is even internal emotional significance to the process of sharing:
When I go on a trip and want to share a photo with close friends, I will take several photos, and send slightly different ones, with a slightly different message, to each one. I do not forward them one shot with one caption, even though they might not know the difference. I want to take a moment to actual write a message to each person, even if it’s basically identical to the person before me.
They deserve my time, and if they don’t it’s not worth sharing.
Speak up for authenticity
Many have already pointed this out – AI is stepping in and doing all the mediocrity for us. The only space that remains is excellence – creating truely out of the box solutions, creating things that are deep reflections of ourselves, which ironically often includes imperfections.
(I’m not even talking about the self-reinforcing loop that happens when you go to AI for validation. I speak here of AI as a tool, not a friend/psychologist/expert. That’s actually a far more dangerous area.
I am committed to never using AI to create a social media post. Maybe an SEO post if I need to technically appease the Google Robot Gods. But for any post meant for a human to read, I want them to encounter me.
We already have labels for highly convincing things like photos and videos to help identify them as AI, and this is important.
I propose the inverse, for what it’s worth (and I know it can be instantly abused): that we label our work as not AI generated, when it isn’t.
That we point out to others, and maybe to ourselves, that we have spent the time and effort to create, to share part of our inner world on a virtual page.
I wrote this blog post entirely by hand. One keystroke at a time, like our ancestors intended.
