I know I'm kinda late to the party with this topic, but as literally everything I do can also be made by some AIs, I feel the need to justify why I'm choosing to not use them to assist me when I do writing, when I create assets for a game, when I do relatively tedious coding, and the such.
Just in case you're not up to speed, there has been a ton of companies that have started offering AIs that assist or replace people for certain tasks in the 2020s. They can be free to some degree, in fact famously, OpenAI has always been operating at a loss and will still be like that for a few years. They're the guys behind ChatGPT, and DALL-E as well back when AI-generated images were a novelty.
So, why not use free tools when they're literally free?
Ecology and the feeling of guilt
Firstly, I do have to admit I'm not super ecological, like, I do take some actions to avoid leaving a carbon footprint too big, but I can't say I'm taking all the actions that I could. I try to afford myself some breathing room in that sense, to strike a good enough balance with my comfort. If I truly wanted to avoid leaving a footprint, I would just kill myself, ya know? By the way, ever listened to "Vitamins" by Mili & World's End Girlfriend?
Anyway, climate change has definitely been weighing on my mind since my teenage years. One of the many, many factors that virtually contributes to its worsening has been the increase in the computational power that is required to run just about everything that is around us.
By the existence of this very website (and my other websites!) as well as the hosting of the video games I make on itch.io and Game Jolt, I do actively contribute to that increase, right! However, I'd really like not to add to it by using generative AI.
Generative AI has quite the impact on the environment, there are articles being written about it since at least 2020, so it's fair to say I've been regularly reminded that regarding ecology, using that kind of tools is fairly inefficient.
With all that in mind, I definitely have an aversion when it comes to generative AI. When I think about using it, I just have a little voice asking me, "But what about the climate?", and that just makes me feel bad enough, I guess.
Companies and the feeling of dependency
It's virtually impossible to run such tools on my device, and I mean running them directly, without a website or an account or an internet connection or anything like that, very much like a normal software you'd run on your computer. It's mostly because of the high specs requirements.
Let us assume I find a generative AI that I like, it's (naturally) behind some website owned by a business. They have a free plan that works just fine, until it doesn't. There are actually several ways this can happen, and it did happen for some people.
A recent example would be when GPT-5 released, as when that happened, they made GPT-4 unavailable, and it's important to keep in mind it's not because it's a new version that it's necessarily better. I remember people complaining about GPT-5's "different personality", as in, GPT-5 needed to be handled differently from GPT-4.
As a service offered through a website, it is also subject to issues unrelated to the fact that it's generative AI, like removing free plans or making them much more restrictive, or increasing prices if access to those tools is gated behind a monthly or yearly subscription. It just... sucks to not own the damn tool.
Creativity and the feeling of lack of control
I just don't feel like I'm in control when the tool generates something, and I can't be one hundred percent satisfied by the result it gives me. Furthermore, achieving that result is hardly satisfying.
Something I'm definitely bad at is drawing, I just can't draw for shit, so you'd think it'd be nice for me to use such things to create assets for my games, but nope! At least at the time of writing this, it's fairly obvious when something is generated by AI, and that fact alone is discouraging enough if you want an art style both not used in thousands of other projects and cohesive.
When it comes to coding, even for repetitive tasks I just don't want generative AI. Putting aside my perspective that code is art, spending time writing code no matter how trivial or repetitive that code is, it's a good way to think and see how you can improve your application and even your coding skills!
Code suggestions aren't terrible, but they do take away some of that critical sense. At least for me, it's easier to learn and to avoid certain mistakes if I do see how they're mistakes in the first place compared to a new solution I thought of!
Regarding music, I just like making use of my own style. Like, think of a written signature, I'm sure an AI could generate a very pretty signature, but it'd very much be like all the other pretty signatures used by other people, either because the AI was trained on them or because those people used an AI to make them, probably both honestly. Friendly reminder that generative AIs train themselves on their own stuff.
I do wonder if it's silly of me to think that way, I guess not but maybe I visited too many echo chambers?
Copyright and the feeling of unlawfulness
Every company out there wants data to feed to their generative AIs. When it comes to code, your code on GitHub is used without opt-in at least by Microsoft, and your code on other websites is collected by scrappers, without a care in the world so that solutions like Anubis had to be created to try countering them.
Regarding actual words, famously, Meta pirated millions of books to train their AI on them. Recently as of writing this, OpenAI didn't care about copyright when it came to training on song lyrics. Speaking of songs, companies are also training generative AIs on copyrighted music!
Lowkey, I'm getting tired of adding links, so just trust me or look it up yourself, but of course it's the same thing when it comes to images, let it be photos or artworks, the generative AIs we know are trained on them without the consent of those who made them.
When there are all these issues, does it feel right to support an industry that seems to consider itself above the law? Don't get me wrong, I think very negatively of copyright laws in the first place, but I think I shouldn't involve myself with companies that keep playing dangerous games with the justice, no matter the laws.
That is of course omitting the question of whether or not you own what was created by the AI, or if it's owned by the people behind the AI, or if it's something else entirely, that unknown status is just scary if you want to commercially use them!
Though it's not all bad
Generative AI can be used for good things, mainly when it comes to accessibility and places where creativity is practically nonexistent. For subtitles for example, I think it's fine to use such tools, as creative derivations are very often unwanted and just typing the words you hear can be very tiring, though in my inexperienced opinion, a human should always fully check them.
For translations to a certain extent, those work too if the "certain extent" in question is dirty work where getting the best quality and fidelity is really not a priority, or in other words, for translations for personal use. In no way should they be compared to translations done by professionals, there's just so much knowledge of context and culture that's involved when doing translations, much more than what generative AIs are currently able to do.
Regarding voice generation, that's again good for accessibility, but it's usually too cheap of an alternative compared to voice actors. Talented voice actors can bring amazing deliveries and emotions in the context of a story, just knowing a certain voice actor is involved in a piece of media can make me want to check it out!
With all that having been said, none of that is immune to the aforementioned issues, so I guess it's more like I get why it's used, but I wouldn't want to use them myself. With the possible exception of Whisper for making subtitles, maybe?
Final thoughts
I think I understand why some individuals decide to involve generative AIs into their projects, I guess they want something that doesn't seem too bad instead of stealing, using free assets that have already been seen elsewhere, or paying what may be too much for an uncertain result, though for the latter I'd argue that's already what you're doing if you're paying for a generative AI.
This is gonna sound cheesy, but I think that even if you're incompetent, the result of something you made will still have some charm that the AI thing will not have! But if you're really that incompetent or embarrassed, using free assets (or even paid assets) is really not that bad, there are many out there and the odds are that no one's ever seen the ones you end up choosing!
When there's something to look at or to listen to in a video game for example, I sometimes like to stop and try to think about what went through the artist's mind when they made that something. When I'm aware that there was no artist to begin with, this takes away so much, really.
Really, the point I'm making with this post is that there are few reasons to use generative AI, in addition to the existence of reasons not to use it, so I will personally stick to methods that don't involve it. Maybe you should too!