Notes on the A(I)pocalypse

an essay by Yaa Mensah-King

On a regular day, my morning routine looks a little bit like this: wake up, reply to texts and DMs, get out of bed, brush teeth, shower, get dressed, and then quickly scroll through what was once Twitter while booting up my laptop. This entire routine takes about an hour but in that short amount of time, I’m forced to encounter AI features at least four times. The keyboard on my phone has an AI assistant button, you can't search for anything on Instagram without encountering Meta AI suggestions, Elon Musk desperately wants you to use Grok and my computer insists that I cannot go about my day without using Copilot and it only gets worse from there when I actually start using the internet for school and work. The first time we lost the Internet was to advertisers and I thought that it couldn’t get more soulless than that but I was sorely mistaken and this time around, it might be for good.

Like most people, I understand that ads are an important way for websites to generate revenue but if you are on the internet as much as I am, you don’t want a billion pop-ups every time you click on a link. The solution to not wanting to see ads was of course downloading an ad-blocker extension to your browser. How I wish there was an AI-blocker of sorts because the tech bros have simply gone too far. I remember vividly how between late 2022 -  when programs like ChatGPT launched, and early 2023 - when people started widely adopting them, I’d see tweets once in a while of people sharing screenshots of them using ChatGPT, Midjourney, Dall-E and other similar tools. I was not inspired to go check out any of them because I can read and write for myself and if I really wanted to do art, I could pick up a pencil and learn because that’s where the fun is. I thought I could just ignore it and keep on minding my own AI-free internet business. The existence of this article should tell you that I had no such luck.

I’m a student so I’ve used a lot of educational apps and websites for years now. Websites like Kahoot, Quizlet and RemNote now all have AI-integration that is constantly recommended to you, Google Docs wants me to use whatever Gemini is and I can’t even open a PDF file on my computer without Microsoft’s AI assistant Copilot offering to summarize it for me the second I open the file. You can’t use Canva anymore without AI-generated icons, stickers and assets being recommended to you before you do anything and Pinterest, which I use to look for design ideas, has slowly been infiltrated by AI-generated images. Some of my professors have made lecture slides including AI-generated content and it’s evident many of the posters around campus are now being made using AI.

A part of me wishes I was just being nitpicky, snobbish and just suffering from a general aversion to change when I am disturbed and disgusted by the prevalence of AI in daily life now but unfortunately, it is not. There are real, catastrophic consequences. A large chunk of AI’s environmental impact comes from the immense amount of energy it uses in the processes of training and running various models. Additionally, beneath the technical and technological feat that is AI, there are ever-increasing carbon emissions as well as the production of e-waste. The more passionate among the AI enthusiasts often come to the technology’s defense with this notion that there is no actual evidence that using ChatGPT is burning down a forest, or that seeing what Timothee Chalamet would look like if he was Brazilian can create a desert. While these examples that skeptics use are often hyperbolic they do have a point. AI tools and programs are not conjured out of thin air. There is highly specialized infrastructure built for them and these servers, data centers and the like do have an environmental impact just like everything else. The magnitude of this impact will remain unquantifiable if companies like OpenAI refuse to disclose just how much infrastructure they’ve built. We can speculate, however, as we have an idea of how many people are using AI daily. The increasingly popular ChatGPT has an active user base of around 200 million. How much infrastructure is needed to service all those people efficiently and effectively? It’s also important to remember that ChatGPT is just one program. With every tech company pushing for AI integration even for the most mundane tasks, the number of people using AI increases significantly. It’s then safe to say that there are a lot of physical factors at play here and too much of anything never bodes well for our environment.

There’s also a lot to say and a lot that has been said about the creative aspect of AI or rather the lack thereof. We all know that AI models are trained on large amounts of human work, often without their knowledge. If your data is on the internet, whether that’s your photos, writing, artwork, or music, chances are it’s been fed to some AI and it’s regurgitating a mixture of what it has been fed to someone not willing to put in the effort of learning by themselves. There is nothing exceptional that AI can do that humans cannot. What I write today is influenced by the things I’ve read, seen, and listened to and I can produce something new from that information and it’s still my work because I put the effort in. Yes, I might make errors that a ChatGPT-generated text might not but that’s just part of being human. Those who are widely read can recognize AI-generated text almost immediately because it sounds like nobody and everybody at the same time. Why sound like that when you can just sound like yourself?

I will not claim to know everything but I believe the increased over-reliance on AI is symptomatic of a society in crisis. There is an observable deficiency of human spirit. We’ve relented the creative tasks to machines and I can’t help but wonder why and to what end.  Is it because we’ve become too busy and we can feel ourselves running out of time? If we’re going against the clock here, why are we so hell-bent on using something that’s likely making the clock tick even faster? When we have ostensibly saved so much time using AI, can we be sure that it won’t be too late for our planet and ourselves?