My Issues with AI

Okay, ya got me. These are going to be my problems with generative AI, many of which are backed up by facts, with sources cited, yadda yadda. Before getting started, though, be warned that for the sake of the blog, I’ll be using “generative AI” and “AI” interchangeably. If you want to get into the differences between the two, I suggest you Google it and see how wrong the AI Overview is. Oops! The snark has started already. Let’s get into it.

Festivus for the Rest of Us (Who Care About Originality)

I have a lot of problems with AI and now you’re gonna hear about them! Seriously, I have a lot of problems with it and since it’s been coming up more often in my professional life, I felt compelled to write about it. Writing. What a concept. 

AI’s Potential

First things first: I believe AI has the potential to be a good thing in certain settings. I also understand that folks with learning disabilities could be and are actually using AI to help perform certain tasks. It would be great if AI could help folks, for example, with dyslexia or with comprehension issues. I hope if you’re reading this, you know my heart, and I’m not talking about the genuinely helpful ways AI could come into play in our lives.

Actually Helping People

The best use of AI I’ve seen recently came to my attention in an episode of “Conan O’Brien Needs a Friend,” titled “Jules is Always Right.” Conan interviewed Jules, who’s living with ALS, and his wife, Maria, living as his caretaker. During the interview, Jules used an eye gaze device, explaining that he banked around 300 short statements into the device using his voice before he lost the ability to speak. They also have a podcast called “The Couple Shift” and were able to use some of those recordings to help generate an AI-cloned voice used by the device. 

Tobii, the company that makes the eye gaze device, worked to put the voice recordings into an algorithm so that Jules could use the device in conversation. A tablet is set up in front of Jules and he uses his eye movement to generate words and hold conversation. Thanks to that technology, his voice lives on even now that he’s lost the ability to use it. 

That’s an example of AI I am 100% on board with and I’m also bummed that it seems to be the exception, not the rule. I worry that folks won’t let ethics get in the way of using AI for tasks that traditionally call for originality, thought, and effort when they do have the ability.

Now, onto my issues with AI. 

Potential Copyright Infringement

The most glaring issue with “AI-generated content” like we’re seeing with ChatGPT or the newer “AI Overview” on Google is how it’s sourced. Is aggregating content created by real people who took the time to research and write it while playing by the rules Google laid out ethical? I don’t think so. Is it legal? Maybe not. 

A group of publishers in France took Google to court in April 2024 over an allegation that it had trained its Bard AI by using copyrighted news articles. The Autorité de la concurrence, an independent administration regulator, fined Google, along with its parent company and two subsidiaries €250 million ($271 million), finding that Google was in breach of an agreement it reached two years prior. That agreement said that Google would enter good-faith negotiations over paying news sources for content. 

AI is Derivative

Recently, Ana Gasteyere was a guest on the “Las Culturistas” podcast. She and hosts, Bowen Yang and Matt Rogers got into AI and the whole conversation is worth listening to (the conversation is kicked off with Matt’s “I Don’t Think So, Honey”). Bowen says that AI will always be derivative, by definition and it’s true. It’s essentially repackaging/copying other work to “create” something out of it. 

Homogenized Art Sucks

By trade, I’m a writer and web designer. At heart, I’m an artist. I’ve written and sketched art since I can remember. I love creating something from nothing—looking at an empty page and filling it with ideas. As I said before, I know not everyone can do that but that’s why those of us who can are here! AI art is just software creating a Frankenstein’s monster out of real artists’ work. In doing that, we’re losing unique points of view. We end up with a bunch of homogenized nonsense with no real point of view and that’s so boring. It’s the shiplap of art. 

It’s Often Obvious

I can clock something that was created with AI in an instant. It’s especially obvious when I get a document or email from someone I know well enough to know how they speak in real life (here’s your reminder to write as you speak). Sometimes the AI is so bad that I feel like I just walked into a room where a dog is tearing a book apart. Pieces of pages float to the floor and someone runs in, decides to tape the book back together without paying attention to the sentences, words, or chapters, and says, “I fixed it.” That’s what AI looks like to those of us who know better.

If I allow myself to get existential about it, I imagine “AI-generated content” as a snake. Those of us publishing our creativity online stop doing that and the snake circles back and eats itself, having run out of content to aggregate. Will that realistically happen in our lifetimes? Probably not, but it’s nice to 

The Environmental Impact of AI

“AI is digital so there’s no environmental impact, right?” Wrong.

The process of building AI models can eat up thousands of megawatt-hours of electricity. It involves chip manufacturing, training, supply chains, etc., and can leave carbon emissions equal to that of hundreds of households. On top of that, data centers use water for cooling. With the recent generative AI services, tech companies have increased their water usage exponentially while we have freshwater scarcity worldwide. Yay.

Here’s where my ethical issue with AI gets deeper. Did we need that AI Coca-Cola commercial? Does anyone need to use AI to write two sentences in an email to a co-worker? Did I need to get on my high horse here? The world may never know.

I’m a tree-hugger and make a conscious effort daily to do my part to reduce my carbon footprint. Maybe if we all imagine buckets of water being wasted on a few sentences or an ugly design, we’ll see a change. Ugh. I know that’s not going to stop it, even as I write this.

Nothing makes me feel like more of a Luddite than witnessing so many people fall back lazily into using AI and either brushing it off as a convenience, acting as if it’s a total win for everyone involved, or never thinking about the repercussions at all.

Wrapping it Up

In conclusion, pay real people for their real work. Writers, designers, photographers, painters, and other artists actually give a shit about the work and deserve to be compensated fairly. Art is good. Stealing art is bad. 

Fin.