Issue #11: Generative AI's environmental impact

If you’ve been on the internet in the past few years, you have most likely heard, or even used (whether voluntarily or not) artificial intelligence (AI). But what is it? And why is it so bad for the environment?

☛ The definition & the difference between traditional and genAI

According to the Merriam-Webster dictionary, the definition of “artificial intelligence” is as follows:

Artificial intelligence: the capability of computer systems or algorithms to imitate intelligent human behavior.

And in theory, that sounds really good, right? Imagine all of the scientific progress that could be done with AI that would’ve otherwise taken DECADES, that was suddenly done and finished by an AI algorithm.

You might be wondering: “What’s then so bad about AI? You just said AI is great for scientific progress." Well, yeah. But this is AI in the traditional sense, and what we’re going to talk about today is generative AI (which is, in itself, only a subset of the AI we just mentioned). GenAI itself isn’t that bad, but the way it gets used is the real issue, and we’ll get to that later.

The difference between ‘traditional’ and generative AI is that for whatever job it’s employed in, traditional AI analyzes already preexisting data (e.g. a trainer in a Pokémon game will be able to choose between what moves should its Pokémon use, swap out the current Pokémon in battle, or use an item, what it was programmed to do), while generative AI creates entirely new data (the Pokémon trainer does a backflip and kicks you in the nuts, what it was NOT programmed to do). Though, the content generated by genAI is based on existing data that’s fed to it, that most of the time is copyright protected material, or there was no consent between the original owner and the company who fed their work to AI.

☛ Unethical model training & environmental impact

Do you understand where I’m trying to get to? AI companies don’t care whether the data they’re feeding to their model is copyright protected or not, and they surely don’t ask for consent when they feed smaller creators’ stuff to it. That leads to a ridiculously big amount of AI generated slop EVERYWHERE! And that generated content is mostly (if not all) a big sign on the AI company that says “I am an asshole because I allowed this shit to happen and I should be sued and go bankrupt".

That isn’t the only problem, though. Generating something, whether it’s an image, text or whatever, isn’t free. I’m not talking about a subscription plan or something, but about an environmental footprint, because AI centers waste a SHIT TON of water.

To generate a 100-word email, ChatGPT uses half a liter of water (~ 17 fl oz), and the maximum amount of words it can generate is ~3000 words per response, making it able to waste 15 liters of water at once, when that water could be used for something more useful, especially because our supply of DRINKABLE water is not endless.

☛ AI girlfriends & the chatbot (loneliness) epidemic

Chatbots aren’t something new. Stuff like that has existed since the ‘90s, but talking to bots like Cleverbot doesn’t have the same impact on us like modern chatbots like the aforementioned ChatGPT or character.ai or whatever else. Some people today prefer having a “partner” that’s a chatbot rather than a real human connection, and this phenomenon is surely not helping with this loneliness epidemic that has been spreading around recently.

Only talking to a chatbot (from my experience, I was a very lonely 14yo once) builds a sort of bubble around you that’s difficult to pop once you want to interact with actual people, and look “normal for once”, but instead, being in that bubble made you forget all of the other aspects of human interactions that isn’t speaking (or well, typing) like body language, to make an example.

Maybe, as you're reading this, some of you still believes that “AI isn’t THAT bad”, you need to hear that this same isolation due to chatbot usage has led some kids to kill themselves, and while it wasn’t fully the chatbot’s fault, it played a big role in these tragedies, as serious red flags were not picked up (or fully ignored) by the bot’s “safety filter”. I won’t talk about these poor kids in this entry (otherwise this’ll get too long) but there are plenty of other articles online that I think you should check out.

☛ Pros (are there any?) and cons

People can argue that AI eliminates human error, it can make decisions with no biases, and it can take on repetitive jobs pretty easily. But I think that its cons make its pros unimportant, many of which we have discussed already. AI is so uncreative and unoriginal, takes away jobs and is expensive not only in terms of cost but also ecologically, and it’s surprising how much companies are willing to do just to have a bot that goes beep boop and actually does nothing.

☛ AI implemented in schools and education

This whole issue was inspired by a teacher of mine assigning us to generate a presentation with some weird website to study from. Thankfully not all my teachers are like that, but shouldn’t a teacher be the first to encourage us to use our brains?

If you begin using AI to “study”, not only you’re studying off stuff that’s most likely incorrect, but you’ll also forget to do stuff on your own, and when you get off of it, you struggle to relearn how to study on your own. So, don’t use it in the first place and if you do use it already, do yourself a favor and unlearn how to use it. I know I just said that it’s going to be hard, but it’s going to be worth it in the end, too.

I’ve heard people discuss other stuff where AI could be implemented, and one of them was AI presidents, and to answer that I quote an IBM slide from 1979 that you’ve probably saw already somewhere else: “A computer can never be held accountable, therefore a computer must never make a management decision”, and I’ll leave you (metaphorically) with that.

☛ Is it worth it?

If you’re still here, you’ll know that my opinion is going to be “absolutely not”. We don’t need AI, everyone was okay before it became what it is today, and if anything, I think that AI is ruining our lives. After all, it’s only really good at guessing words.

If you use AI, let it go.

-Dam