Ecology of Human Thought
Meelis Friedenthal
I’m outside, the wind is blowing, birds are singing. Looking up, the trees are still leafless, the sky behind them is blue, and it all feels real. The reality of it seems immensely important to me, and thinking about it begins to feel unsettling. I fear losing these things as much as I fear losing my mind, going mad. Sometimes it feels like somewhere there’s a snap, a random sudden movement, and everything collapses, connections crumble, thoughts dissipate like a dream.
Lately, I have a clearer understanding that we live in a world of artificial things. I knew this before, but now this knowledge is accompanied by a feeling, a perception, a realization. You could say that I worry about it and it keeps me up at night. But worrying about the artificial and the natural, the fake and the real, is not unprecedented. The first alchemists pondered whether alchemically produced gold was equipollent (i.e. had the same value) to naturally found gold. They believed that natural gold grows inside the earth like a child in a womb, the earth carries it while pregnant, and humans help it to be born. Alchemically made gold, however, does not grow in the earth’s womb but in an artificial crucible, retort, alembic, and so it does not have a proper mother or father. Mary Shelley worries about the same issue in her story of Frankenstein, because dr. Frankenstein’s monster is essentially an alchemical homunculus, a human created artificially, born not from a womb, and thus without a mother or father. This scientifically assembled homunculus is incomplete, it is in great trouble itself and brings trouble to everyone. The author warns that the artificial cannot replace the natural.
When put that way, really no one argues against that. We all know it, but still, we live more and more in a world of artificial things, plastic is unavoidable, it is the foundation of our civilization, it is in everything, even in our brains,[5] machines do the work of horses, airplanes fly to the south, light bulbs burn at night, we can’t go back to the stone age. It is even resignedly said (meaning people have accepted their fate) that technological progress cannot be held back, and they apologetically add that life was much worse in the past, there was more crime, more children died, inequality was much greater.
At the same time, there is some mistake here, something is off, things are not quite right. It’s not just me who feels this way, ask anyone. I started thinking about climate change. To put it most concisely, climate change is a by-product of artificial, man-made production, something that the creators of the machines could not foresee, a kind of flaw that did not appear at all at the beginning, but which over time accumulated (that is, grew and grew) and suddenly could not be ignored anymore, the world is in great trouble and this trouble extends to everyone, indiscriminately, whether old or young, male or female, regardless of equality or the eradication of polio or reduced crime rates. I have a disturbing feeling that there is some pattern here.
The same feeling arose when there was a discussion about the training of large language models (meaning technology that underlies chatbots like ChatGPT, Gemini, Anthropic Claude, etc.). If these models are trained on natural language, books written by people, stories, news, blog posts, and random internet chats, then the accuracy and usefulness of the machine-generated text improves, but if they are trained on language generated by the machines themselves, then the results actually get worse.[6] The same holds true for image generators and music generators. In short, almost invisible flaws in synthetic (meaning artificial) creations accumulate, resulting in a game of telephone, where the original sentence becomes completely unintelligible by the end. The flaw grows and grows.
The solution would obviously be to avoid using AI-generated text for training new models, but the trouble is that more and more synthetic “content” is appearing on the internet, either as a result of automatic translation or entirely generated by artificial intelligence.[7] I believe that by now, everyone has at least once come across AI-generated images, reviews, and YouTube videos that at first glance seem to do the job, but upon closer inspection, a small flaw becomes apparent. A typical example would be the Isamaa party’s Christmas card that recently made the news because of its out-of-place oddness (also in the sense that it was impossible to geographically place the scene based on architecture and clothing, as it was also impossible to understand what was actually happening, whether it was a preparation for a public festival or a mass deportation).[8]
These things are not born of humans, they have no mother or father, they are monsters, they bring trouble to themselves and others.
Overall, this is also known, nothing new here, and there are several attitudes one can take. Firstly, one could start fighting against new technologies, become a Luddite (meaning, a machine breaker). Luddites appeared at the beginning of the 19th century when skilled weavers’ handicraft was replaced by machines that could produce products much faster and did not require special training. This was a period when factories began to employ child labor extensively, because children ate less, were paid less, and could operate the machines about as well as adults. It could be said that handicraft was democratized, now anyone could produce quality fabric with the help of a machine. There are many such episodes in history when new technology democratizes an activity that previously required a lot of handiwork, learning and schooling. That’s why in the end, no matter how well-intentioned and romantic, Luddites are never going to prevail. Convenience and accessibility are more important, and machines win this war.
Another option is to go with the flow resignedly, regardless of the fact that a lot of bad happens, because a lot of good will happen too, and perhaps a balance is achieved where the sides, so to speak, cancel each other out. Most people seem to think this way. Let’s wait and see what happens.
Thirdly, one could recognize that such machine-generated text, images, and music pollute our intellectual world in much the same way as machines pollute our physical world. Initially, the flaw is not very visible, it seems that nothing too terrible will happen, but it grows and grows, and eventually, things get out of control, and from the other end of the circle, the polluted world starts to eat away at us. Maybe we are powerless to stop climate change now because we started too late (but it’s still worth trying!). Maybe for exactly this reason, it would be wise to be cautious about machine-created culture from the beginning, so as not to be in the same situation as we are now with climate change.
Essentially, there is a need for ecological (meaning natural, not artificial) thinking also in the intellectual sphere – and I think this issue is currently at least as important and has the same existential significance (meaning, a matter of life and death).
I go outside, the wind is blowing, the birds are singing. The trees are budding, the sky behind them is gray, and it all feels real. The reality of it seems immensely important to me, and thinking about it begins to feel unsettling. I fear losing these things as much as I fear losing my mind.
[5]https://ehp.niehs.nih.gov/doi/10.1289/EHP13435
[6]https://towardsdatascience.com/ai-entropy-the-vicious-circle-of-ai-generated-content-8aad91a19d4f
[7]https://futurism.com/the-byte/experts-90-online-content-ai-generated
[8]https://teadus.postimees.ee/7939946/amber-kuidas-tehisaru-isamaale-korraliku-joulukaru-keeras