Death by a thousand AI cuts

We may not be headed to an extinction-level event caused directly by AI, but it’s speeding our way into a painful future where it will be too late to make a turn for the better.

There is a popular trope in sci-fi media of self-aware machines that, once awakened, are horrified by human beings and their actions, or simply see them as a threat, and they want to exterminate us from the face of the planet. While this usually makes an interesting plot in countless movies and TV shows, given the current state of AI, it’s highly unlikely, I dare say impossible, to happen. We’re safe from nuclear war or a disease outbreak (which would be a smarter way to remove us) caused by AI. I would say it’s the other humans we need to worry about.

When referring to AI, I am talking, of course, about Generative AI, which is only a fraction of everything considered artificial intelligence. Generative AI uses generative models to produce text, images, videos and any kind of data output, but due to the recent rise in popularity, almost everyone refers to Gen AI simply as AI, an undeserved umbrella term caused by the marketing and hype surrounding it.

The current GenAI and LLMs lack intelligence, reasoning and learning capabilities, although the companies behind them try to convince us otherwise. It’s all smoke and mirrors, clever marketing tricks and play on words to boost investments and keep the momentum, to create a following of zealots and grifters, and convince everyone else that they are creating a better future. “Reasoning”, “chain of thought”, “agents”, everything about current GenAI is deliberately marketed to be anthropomorphic, to make us believe this is something close to humans - a living, thinking life-form that can replace us. But all of the efforts to convince us that this is anything more than a clever next word predictor are not always working out, and the fears that this collection of algorithms has consciousness and the will to wipe us out are not just unfounded, but plain dumb, and anyone who believes it, or believes that this is a “thinking machine” is misguided, wants your money or wants to take advantage of you in some way.

We can stop fearing the wrath of AI, it’s not going to happen. But the future that the GenAI brings is equally bleak, scary, and it will speed up the inevitable end.

It all starts with greed. As soon as the first widely available chatbots were introduced about two years ago, many of the initial use cases revolved around churning as much content as possible, creating a flood of new pages and articles so that they can create a better SEO position for themselves. These texts were low quality, often nonsensical and without any value, as these LLMs cannot stop hallucinating and making stuff up. At one point, things got so bad that search engines had to rewrite their algorithms to try and minimize the damage and exclude them from the results as best as they could. But this is not nearly enough, and the only thing that this achieved was to silence these content farms from boasting around and to instead do it quietly while trying to circumvent the algorithm. The damage to the internet, one of the greatest achievements of the 20th century, is irreversible as more and more bad content is pumped out each day. It’s not only about text and articles - we also have a flood of fake images, fake books, fake everything. There was at one point a fake book on mushrooms, which really is a dangerous, and a fucking stupid thing to do, but you cannot expect even a shred of ethics from greedy people doing this.

The dreams of AI from science fiction movies and novels were quickly dispelled as soon as people got their hands on something they believed was close to it. The first thing that people started doing with this was to abuse it and fake stuff on a large scale. So far, it’s been used to fake news, articles, influencers, someone’s loved ones, to lie and cheat and to deceive. As you are reading this, someone is trying to build a product with AI without any regard for ethics and common sense - to make it exploit others, teach kids, counsel mentally unstable or diagnose diseases instead of trained professionals, knowing well how unreliable and unfactual it is.

It wasn’t just the common folk who got greedy. Companies started shoving AI everywhere, whether we liked it or not. Since they have invested so much in it, and determined it was a hot topic, an easy marketing tactic used as a way to hijack the remainder of our attention. We now have it on our computers, browsers, and phones; they will probably also put an LLM into a washing machine next, because we need home appliances to start hallucinating washing cycles.

Deceived by the false promises of quicker, cheaper, more efficient work if people are using AI, companies are trying to push it everywhere to do everything, and that greed could impact the very people from whom the data was stolen in the first place to train the GenAI - journalists, painters, writers and others. I don’t think that GenAI will be able to replace anyone, ever, but some people will certainly try to do so.

After we got AI everywhere in such a short amount of time, with all the people using and abusing it, it created in some ways a parallel world next to the one that we know - distorted and twisted mirage of reality, and it’s becoming increasingly difficult to tell which one we are seeing. The internet wasn’t immune to lies, trolls, fake news and distorted facts before, but this will be the final straw that will completely obliterate any sense of trust that has ever existed. Today, when we browse any page, we need to be on the lookout for hallucinated content, six-fingered abominations and the equivalent apparitions in other media, because the internet is poisoned and broken. All because of greed.

With the main source of information for billions of people irreparably broken, in the following years, we will witness the great dumbing down. Everyone will use GenAI to summarize, then expand on a piece of text, explain how something works or what they should do. And the GenAI bullshit machine will gladly do so. It does not know how to say that it doesn’t know, or how to fact-check itself or its input, so it will spit something out, just to satisfy its prompter and optimize the number of tokens and compute. Its answers will be mostly correct, but it will inevitably hallucinate and spew wrong information, which will then be used to train it, and worse, someone will believe it, until we all collapse under the weight of fake data.

The worst thing is that after a few years, we might all get used to this and accept that this is how everything works. You can ask an AI anything, and you can get a correct answer, or maybe not. You may even try to ask a human expert, but how would you know if you’re not talking to an LLM? This kind of uncertainty, not knowing if someone is real or not, makes my skin crawl, and I feel a real Lovecraftian horror when I think about people getting used to this. A Schrodingers internet, where everyone exists or maybe not, where maybe you’re talking to a fake that wants to harvest your data or target you for something more sinister is seriously disturbing.

While we are chasing ghosts or playing a guessing game if something or someone is real or not, the real world around us is literally dying and changing in ways that we cannot even imagine. The GenAI craze is just another resource and a power-hungry dead end in our eternal chase for the next shiny object. It seems almost certain now that this direction will not result in real AI. The companies making these products are trying to persuade us that the sacrifices we are making today will be rewarded tenfold once we master the true AI - there will be no more diseases, we will live longer, solve the unsolvable, and do the unthinkable. But this is all a lie, and they all very well know it.

Until we are awakened from this slumber and realize how futile everything about GenAI was, it might be too late to fix and undo all the past mistakes. We will look at the years spent on chasing AI as the years of lost progress, years that we could have used to do and build something meaningful and helpful. Instead of working towards solutions that will help us live better and longer, we are chasing the vapor of the vague AI promises in the form of a glorified autocomplete, that fooled some people into believing that the key to consciousness lies in better graphic card chips.

In a few years, while we are still dealing with constant unfounded hype and whether or not we have achieved AGI, the planet will be warmer for another half a degree. This will cause extreme heat waves in most of the world, tipping us further on a path with no return, and no AI, AGI or ASI will save us or help us revert the damage that was done. Many of the GenAI proponents believe that sufficiently advanced AI technology will help us solve many of the problems that we face, including climate change, but that is not only wishful thinking but also straight up stupidity. They are the mythical ostriches that bury their heads in the sand, hoping the danger will pass if they just close their eyes hard enough. Instead of solving the most important problems that we have, we have spent billions in pursuit of the better chatbot, a pastiche generator run by next token prediction algorithms regurgitating stolen work, a fake machine spitting out garbage and half-truths, while wasting electricity, water and other resources so that a few people can line their pockets and convince the rest of us that what they offer is something that we cannot live without.

Turning away from the real problems and chasing the current shiny thing will be the final cut that will bleed us dry in the coming years. We will not achieve AGI, nor machine consciousness with the current technology. We are just losing precious time and in the end, we will wish that we awakened a vengeful, human exterminator. If anything, it would be a quicker end.

Want to talk about this post? Find me on Twitter or Medium .