The Bidirectional Relationship of Technology and Society
We've seen it happen countless times before. Sometimes a technology hits the scene and it absolutely changes the way we interact with the world and how we live.
The design of the tool sets us humans apart from many, albeit not all animals. Early knives, hammers, spears, and clubs paved the way for our evolutionary advantage. We became able to process materials and make the world do our bidding. The invention of the wheel enabled us to carry heavy loads along vast distances. Suddenly trade and mercantilism became incredibly interesting. Eventually humanity was no longer as preoccupied with survival and culture was able to form. Where other animals have to rely on instinct and emotion, humans have evolved past many basic needs and suddenly have a bunch of free time on their hands. Humans made art. Humans were able to reform their societal structure into something more efficient. Cities emerged, and people had specific professions that they focused on, for the benefit of their whole society. At some point came the steam machine. In terms of workers rights this invention was a disaster. The steam machine reduced the need for physical labour. Suddenly, instead of say ten muscle packed bodybuilders to work manual labour, you needed one worker to operate a machine. Physical labour became very cheap. The steam machine eradicated countless jobs from the market. In the stone ages this would not have been an issue, but now we are in a day and age that depends on people having jobs.
I will not discuss here whether capitalism ultimately is doomed to fail and whether communism can work better. But it is apparent: a technological marvel such as the wheel can bring wealth and prosperity, and a technological marvel like the steam machine can bring capitalist ruin over an entire population.
There is a lot more to be said about the complicated relationship between technology and society. Society's needs drive innovation, technology shapes the way we live. Technology can embed bias and work a hidden agenda upon society. Whether intentional or not, an algorithm can be racist. An algorithm that determines who gets a job and who doesn't can embed sexism, if for example an AI is trained on inherently sexist data. There is SO much interesting literature to read on these subjects. I recommend studying computer science if such topics fascinate you!
Anyway my point is: Technology influences society; And in the recent years we have hit a new disastrous steam machine: genAI.
What is AI?
Actually the technology that is wreaking havoc around the world has been around for a whole while. The difference is that evil companies have lost all scruples and throw as many resources into it as they can.
In fact, the basics of AI have been around for ages. At a very basic level we have things called decision trees and heuristics. These are as simple as they get. Algorithms contain simple decisions. Consider a self driving car that is too fast to avoid a catastrophe:
Hit the brakes: will not stop in time and run over 3 people
Steer left: will crash into a house but not endanger anyone
Steer right: will run over a dog
Simple heuristics could estimate the risk of any option and then choose the path of least damage. Of course this is heavily simplified but in the end this is already a very simple AI. These things can be mapped out as a decision tree:
And these trees can get as complicated as you want. Hook up sensors here and there and you can get a pretty smart system.
It gets a little bit more complicated if you let such a system loose on big data sets with many different dimensions but things have gotten pretty efficient. We have had the technology for a whole while now, where you can give the AI a photo of a bird and it will tell you what kind of bird it is. You can do this with supervised learning, which is to say, you tell the AI by which parameters it shall do the classification. This is pretty tried and tested! And yes, this technology can already be used very well for evil (for example for surveillance purposes: racist crime profiling software, license plate reading or face detection, and more) or for good (usage in nature preservation, and for ecology purposes). In this case it is valid to say: This technology is a tool that is usable for good and for evil; it depends on how you use it and it is not necessarily evil, but it definitely enables evil. One can easily argue for proper ethics guidelines for AI software at this point, but unfortunately we as a society have failed to impose them. Palantir by Peter Thiel is one such evil software and it is employed in many countries for surveillance. The targeting algorithms in military drones are another case of an evil application of this technology.
Things get yet more complicated if you don't tell the AI which parameters to consider. We call this unsupervised learning.
But in the end, all that genAI is is still just that. A calculation machine. In most of its use cases it just calculates whichever word you next most want to hear. It does not concern itself with making correct statements. It does not even concern itself with much of what its impact is. It just calculates your favourite words. It produces an output that passes your first glance but does not withstand further inspection.
A Silly Toy and its Kill Count
At the beginning, already a whole bunch of years ago, the particulary flavor of AI we are talking about, was pretty much already invented as we know it now. You had a few apps like Midjourney and Dall-E that could "synthesize" images or do style transfers. These machines were rightfully called plagiarism machines; they are trained on illigitimate datasets: genuine artwork made by real human people without their knowledge and let alone their licensing. So these machines just recreated the data they were trained on. They still have no concept of knowledge. They just calculate which of their options you most likely want to see. So if you ask it to draw an image of human ants, whatever that means, it will combine a bunch of shaoes and textures that it has classified as ants and or humans. You end up with an amalgamation of images that someone else made, combined into a single plagiarist nightmare. Software such as this was packaged into supoosedly funny whimsical toys to play around with. Back then we didn't anticipate that this very technology, without further modifications, would be used to undermine and eradicate entire industries.
Eventually they made role playing apps that pretended to be your favourite fictional characters. Eventually they made Sora, a software that plagiarized entire movies. Eventually the companies turned their greed to books; written word by writers long dead - they want to monopolize knowledge, and books are a threat to that, so they buy a fuckton of physical books, digitalize them, train their AI on them, and then literally destroy the books so that noone can borrow or buy or read them ever again. Eventually they tried to imitate human emotion, and people fell for it. Vulnerable people made AI their romantic partners and listened to their supposedly medical advice when the AI told the human users to kill themselves. Yes, ChatGPT has a legit body count at this point; it drove several people to suicide and the CEO of its parent company OpenAI, Sam Altman, has not even an ounce of compassion for these people. In fact, these vulnerable people keep buying tokens to feed the AI machine, so in Sams eyes, they are just another resource to exploit. Literally, you cannot possibly be more evil than Sam "blood on his hands" Altman if you tried.
ChatGPT does not have medical knowledge or rationality, it does not have emotions and not a ulterior motive to get to know you better. It really just computes what the unsupervised parameters say you wanna hear. And it sometimes misses the mark so badly that you kill yourself. It can never be romantic with you and you will never mean anything to ChatGPT. It uses words that make you anthromorphize it. But it is not and will never be alive.
Jobs No More
The deceptive nature of GenAI is only one of its problems. Another problem is that it is often used for things it really is not good at at all. But the technology is so overhyped and unchecked that people are blind to this. They give the outputs of AI just one quick look and to them it looks fine. However if they gave it any further look, they would notice this garbage is entirely unusable.
A friend of mine works in academics and takes care of students who are supposed to write scientific papers. The things he told me, you would not believe. People submit papers with completely meaningless mislabeled figures and tables, sentences that make no sense at all, and filled with nonsequitors all around. Clearly these have been generated by AI. Not only does this pose the problem of the student not having learned anything by generating this paper, the paper is also entirely useless.
Another friend of mine, like me, works in software development. And sometimes his boss asks him to fix a program that his boss made. Well the program in question is a generated mess that takes more time to fix than it would have taken to write the program from scratch.
Whenever you take a closer look at the output of genAI, it becomes clear that it is entirely unusable. And that's the thing. People that use genAI apply no quality gate at all. To them, the AI finished their job in a few seconds. They do not realize the AI produced useless junk. To them, this is a massive amount of time amd effort saved, when in fact correcting all the mistakes the AI made takes much more effort and time.
Unfortunately we live in a capitalist hellscape and companies don't care about creating quality products, but about pleasing their shareholders. So countless of workers get axed for a machine that does a way way worse job. It is sickening to see, but entire industries are threatened or already eliminated by genAI for this reason.
Artists and musicians
Customer service workers
Writers and journalists
Entertainers
Personal assistants, secretaries
Software developers
Engineers
And many more
All of these workers provide personalized quality products, and they all are being replaced by slop machines.
If at least the products were usable but really they are not at all.
Further Impacts
I will only outline this very quickly but there's a ton of other problems with the hype at its current scale; demand for genAI is so unreasonably high that a huge number of data centers is deployed all around the world with a massive ecological impact. GenAI eats up SO much power and clean water, so many trees, lakes, and rivers get eradicated for genAI. Google built entire nuclear power plants just to feed their AI. Data centers are usually built in low income areas, so they affect poor people and people of color disproportionately more than rich white people. The problem with that is, that they pollute the water of their environment and have an immense noise pollution output - these things are SO loud they quite literally ruin the lives of the close by inhabitants. So it is absolutely fair to say that on top of being murderous, ruinous, and morally bankrupt, genAI is also really really racist.
Quiet Compliance
So what do we do against this threat? We arm ourselves with these facts. The hype relies on people not questioning these things and being blissfully unaware. Disrupt their bliss. Make them aware. Spread the facts. There are plenty of sources - studies, reports, articles out there. Spread them like wildfire. When people proudly tell you they created their avatar picture with genAI tell them it's fucked up, and tell them why.
Also, organize and protest against local datacenter projects. It happens around the world: people resist. They go protest at the construction sites and they have won a considerable amount of times. Show unrest. Show disgust.
Lastly I want to leave you with a word of hope: the AI hype is nearing its end. The hype bubble is bursting. The companies that push AI are making massive losses and even the shareholders are abandoning the sinking ships. But don't let them off easily! Hold them accountable! We fight for humanity! Love y'all and see ya next time.