I am becoming increasingly convinced that AI is not only a gargantuan bubble, inflated solely by humanity’s disproportionately high self-esteem and a mindless drive for consumption and money, but also a tool that will lead to the degradation of an entire generation.
(1) What raised my suspicion during the first wave of Large Language Models (LLMs) rise in popularity was the ease with which people began calling them "generative artificial intelligence" at first, and then simply "AI"—Artificial Intelligence.
As a linguist, I dislike it when words and terminology are used incorrectly, for I immediately see malicious intent behind it. Manipulation. You see, the term "artificial intelligence" used to be reserved for a SOMETHING that can come into existence when a machine realises it is a machine — meaning, it acquires consciousness. This concept, of course, is ideological in a philosophical sense (no one has figured out yet what consciousness actually is), and sci-fi in a practical sense. That is, everyone understood that if this were to happen in reality, it wouldn't be anytime soon, and it would mark a massive upheaval in virtually everything, including our understanding of who we are as a human race and where we are heading.
In sci-fi universes (The Terminator as an example, The Matrix works too), the emergence of artificial intelligence usually signaled the rise of the machines and the enslavement of humans, or led to a prolonged war. Somehow, several generations of writers, screenwriters, directors, and musicians were certain that machines possessing their own consciousness (and, therefore, their own will) would inevitably have intentions that clashed (or would eventually clash — à la Asimov) with the ones of humanity.
A classic read on this topic is the legendary short story by American sci-fi writer Harlan Ellison, "I Have No Mouth, and I Must Scream." No need to thank me, enjoy the read. Just keep children away from the screens; things go very badly there (plot-wise), and it is very graphic.
And then tech companies realised they needed to sell their ChatGPs and Claudes to the general public and to other tech firms, so they opted for a linguistic compromise without waiting for their creations to acquire consciousness (since, as we mentioned, things aren't fully clear on that front yet). For the sake of a catchy phrasing and more direct, impactful marketing, they labeled it all 'Artificial Intelligence', or simply 'AI'. Doing it prematurely, so to speak. To make it easier to announce innovations and convince the public that you simply can't get by without 'Artificial Intelligence' anymore.
(2) The second thing that has been bothering me for a long time is that Web 2.0 introduced a severe decline in the quality of life for its most active users, primarily due to the rise of social networks and related tools, whose very existence and development are tied to the commercialisation of human attention. To put it simply — the internet has come to closely resemble a casino. All popular news, entertainment platforms, and social networks make more money the more time you, as a user, spend viewing and reacting to the content they feed you.
As a result, for two decades now, tech giants have been paying millions of dollars to teams of top specialists to design online products in a way that fosters user addiction, making it impossible not to use them. They rely heavily on the exact same approaches used for decades prior by casino gambling developers. Technology has leaped forward, but people — and their weaknesses — haven't changed, and humans can be exploited using the same methods as before. This is exactly what TikTok and Instagram do, helpfully curating the precise type of content that has the highest chance of keeping you from closing the screen or the app.
Thus, the internet has turned into a place that must now be used very deliberately and in strictly measured doses — otherwise (as many of us have already realised through our own bitter experience) you risk, at minimum, losing hours per week spent in aimless, mindless, and useless scrolling of social media feeds, and at maximum, a severe addiction that leaves a person unable to spend even a few minutes without their phone.
And let me tell you this: a person who constantly craves a quick dopamine hit, delivered through the screen of an internet-connected device, looks from the outside like an invalid (I am not trying to offend people with disabilities here; I hold them in total respect), or — you guessed it — like an addict.
Both options are terrifying, not just because of their grim imagery, but because of how easily we agreed to let this become a social norm.
(3) That was Web 2.0 — the second iteration of the internet, if you will. It has already led to the appearance of a massive number of disorder terms in our language that simply didn't exist before. What is currently called AI threatens to launch Web 3.0, and if the previous attempt was, to put it mildly, 'so-so', I see absolutely no reason to believe that the next iteration will bring people happiness and prosperity.
And here is the third thing that bothers me. It isn't even the fact that Web 3.0 is already quietly seizing territory (it's hard to say exactly what percentage of texts available on the internet are already written by neural networks, but there is no doubt it will soon hit 99.999...% — your LinkedIn feed already consists almost entirely of them and has completely lost its purpose, even against the backdrop of what it used to be). Rather, it is how meaningless, and often simply harmful and degradation-inducing activity it spawns by its very existence.
It is worth noting that AI does have positive applications outside the scope of what I am talking about — applied AI that enhances pattern recognition in medical diagnostic software built on big data principles and computer vision, for instance. Wonderful. But that makes up 5% of the industry and 0.1% of all discussions on the topic.
What we are actually talking about is how AI is excellent at inventing: (a) ways to retain user attention (remember the casino principle — to foster addiction in a person, you don't need to create meaning, you need to create triggers, and the more meaningless and vivid they are, the better), and (b) making people less happy and more helpless.
With happiness, it’s quite simple: Cal Newport (and before him, a host of writers, philosophers, and psychologists, dating back to Aristotle himself) argued in his book Digital Minimalism that to gain true satisfaction, one must labor, apply effort, join forces with other people, and see the results of their work (preferably in the real, physical world) with the full awareness that they personally had a hand in achieving it, and that they would be difficult to replace.
Nowadays, it reaches ridiculous heights: 'try doing it with AI' has already turned into a sad meme and a joke, because it often results in having to spend more time and effort delegating a task to a program instead of just doing it faster and better yourself. I know that even now many software engineers will back me up on this — they wised up and started realising that writing code with AI is no longer that intriguing, the quality is lower and inconsistent, and to make the code readable by others, let alone catch all the bugs and optimise it, takes just as much time as if they wrote the code the old-fashioned way themselves.
Yet programming is an area where the advantage of using AI was considered the most apparent and indisputable. In other fields the situation is worse, not only because AI is capable of being exponentially better at generating meaningless content aimed at retaining your attention, but because the majority refuses to understand (or accept) this.
Here is one of my favourite examples: about ten years ago, it became popular to make 'summaries' or 'digests' of massive books — for when you don't want to read, but still need to know what happened. Generally, book summaries on Wikipedia quickly occupied this niche, so it never quite turned into a serious business. But then 'AI' came along, and now practically everyone is trying to feed it every email and article just to get a 'summary', or "general overview" of what it says, without spending time reading the actual material.
Do the most active users of AI assistants realise that this function is actually already built into our (and even their) brains, with which they are ALREADY equipped, and for the use of which they don't need to undergo additional training, and it’s called 'skimming'?
I am not being ironic. The Pandora's box has already been opened for research into how constantly turning to AI for every question affects our cognitive abilities. The conclusions, naturally, are discouraging: the use of chatbots is linked to a decline in cognitive metrics such as perseverance (grit, if you will) and reasoning skills.
And we both know that with the advent of social media algorithms curating a personalised news feed of one-sided viewpoints for each separate user just to maximise collection of reactions and make more money (see point 2), humanity’s reasoning skills as a whole ALREADY face massive problems.
(4) People (through businesses and entire industries) stubbornly continue to multiply activities that are the polar opposite of what leads to a long, happy, and meaningful life. Longevity researchers, psychologists, philosophers, writers, doctors — they all figured out long ago that we are happy when we spend quality face-to-face time with family and friends, when we make something tangible with our own hands and reap the fruits of our labor, when we can regularly retreat into solitude to be alone in silence with ourselves and our thoughts, when we move a lot and lead an active lifestyle, spend a lot of time in the nature, and when we show initiative and help others.
Guess yourselves whether the most popular methods of using AI bring us closer to or distance us from the above, and if they distance us, then how quickly it is happening. It is about time to think about the concept of singularity.
(5) Of course, of course, everything comes down to money once again, and how easy it is to commercialise and to profit from AI solutions makes the situation far more dangerous. Do you know why over the last couple of years every top manager, every director, every business owner (we are talking about fields closely tied to digital technology — plumbers and gardeners can relax for now) has uttered that phrase which, at best, provokes a nervous laugh from everyone: 'let's do this using AI'? Because they blindly believe it will either save money or help them make more money.
But will it actually help?
An industry that essentially began its existence with a minor linguistic trick (see point 1) is hardly engineered to deliver on its promises.
(6) What is to be done? Neither my personal resistance nor yours, nor calls of 'come to your senses, people', will work; moreover, lacking a targeted impact, they will only spawn endless arguments. We will just have to watch events unfold — tech bubbles have burst before, entire industries have gone bankrupt, and entire countries have fallen into financial crises lasting for years. Our task is to protect ourselves, to become, in Nassim Taleb’s terms, 'antifragile', so that the problems that will inevitably arise due to the oversaturation and overvaluation of AI do not affect us, or better yet, make us stronger.
Personally, to achieve this, I would not just try to engage in activities where AI has the least influence, but I would go completely against the grain — I would take up something for which people, tired of meaningless content on social networks, are realising a growing need, and I would try to fulfill that need. Successful examples already exist — for instance, businesses related to sports with a social element, like CrossFit or running clubs. And I don't say 'AI' at all — I say don't use it for the sake of using it or for multiplying activities that are contributing to the deterioration of a human race.
In parallel with this, I would strictly monitor how dependent we are on technology as a whole, and actively reduce this dependence. You can start with your smartphone and social media, replacing entertainment and relaxation in front of screens with more active pastimes (especially those where you can gradually improve, see progress, and which require physical involvement). I, for instance, have finally started playing tennis regularly again.
In short, now is not the time to relax. The war against machines is already well underway.
P.S. I showed this text to one of the AI language models. The response: "The text is strong, but don't write that AI is to blame; write that people are to blame."