Reflections on artificial intelligence
A viewpoint on AI, its spread, its nature and its limits.
It will have escaped no one's notice that the subject of artificial intelligence has been one of the most commonly discussed topics in public and private discussions for several years now, and particularly since chatGPT came onto the market.
I felt it was important to discuss this subject in a rigorous way, so as to avoid superficially-constructed opinions, if not simply psittacism of journalistic teasers.
This rather long article has been organized into three parts, the titles of which are quotations I have heard here and there and to which I would like to respond in writing so as not to have to repeat myself indefinitely.
"AI is going to take our jobs."
The most recurrent concern that emerges from most discussions about AI is that of losing one's job to algorithms, which are supposed to produce value equivalent to that produced by workers. With a few people and a good AI, we should be able to do the work of dozens of people, and thus save wages and human resources.
There are several things to understand behind this concern, which is legitimate and already a reality in many companies. It is the same concern that affected craftsmen during the industrial revolution of the 19th century, when sophisticated machines did much the same work as they did, faster, cheaper and with fewer errors.
However, this kind of thinking is a smokescreen that tends to evade the real fundamental issue facing our whole society: that of our economic and political system. The real problem facing workers is not that human beings have been able to produce relatively sophisticated algorithms capable of competing with them in certain fields, but that the spread of these algorithms can be massive because their use maximizes the profits of capitalist companies whose only decision-making cursor is the accumulation of wealth (capitalization).
AI here is just another episode in the history of human enslavement for the benefit of capital, via technologies that remain fairly politically neutral in their basic principles. This episode is certainly spectacular, but no less so than the steam engine or the calculator. We must therefore be careful not to focus our analysis on this new technology, because it doesn't really change the reality of the current system we live in.
I won't go into too much detail here, but this reflection should lead us to consider why we work and produce goods and services, our relationship to quantitative and qualitative production (which, incidentally, shows in some cases that old craftsmanship has still not been overtaken by machines) and our relationship to time and constraints. I will come back to this last point later in the article.
If we unravel the coil far enough, as I do in other articles, this raises real existential questions, where work can have several purposes. However, in a system as totalitarian as globalized capitalism, it tends to have only a commodity function: we trade in what others are willing to buy, and since there will always be someone somewhere to buy something, we are gradually led to trade in anything. First and foremost, however, the act of selling must be motivated by something. We don't sell our time, our arms or our intelligence for nothing.
The additional condition necessary for this omnipresent and omnipotent market system is the incentive to market, to sell and to buy, and the only really effective way of doing this is to create lack absolutely everywhere: housing, food, health, relationships, leisure, etc. This lack can have two origins:
a fundamental lack linked to the essential needs of human beings and largely provoked in the West by coercion (standards and taxes) and opportunism (insider trading): being hungry, being cold, feeling alone, feeling in danger, etc. This artificially exaggerated lack leads us, for example, to constantly look for a job where the real value we produce is equivalent to 200% of our salary, to sell our furniture to be able to pay the property tax on our house, to sell our company shares to be able to bring the workshop where we work up to standard, and so on.
an artificial lack, extended from natural needs in a delirious way by social engineering (advertising and journalism): wanting to drink a coffee at Starbucks and eat a hamburger at MacDonalds, wanting to live in a home where the temperature can be regulated to the nearest tenth of a degree, wanting to have thousands of followers on Instagram and to party every night, ordering fifty cameras connected to a smartphone to monitor a property 24/7.
The other important process to ensure is that these lacks will last as long as possible and therefore need to never be entirely fulfilled, just enough for the market to continue to grow more and more.
Of course, as I often say, getting out of such a systemic grip is no easy matter, and the worry of losing one's job is completely justified by the fact that, in some cases, the risk is real and its implications potentially disastrous in the short term. That is why it is best to work as early as possible to create spaces that escape the grip of this system.
Where possible, a simple answer to the spread of AI in our economy could simply be "no thanks", because the economic design we put in place will be able to afford such authority (enabled by independence). So, for example, we won't look at bread made by a robot in the same way as bread made by a baker. Perhaps the bread will taste the same in both cases, but the implications of its trade within our community will be far greater than simply trying to sell something, because we will have been able to invent a model where the latter is relative and not supreme. In this case, perhaps people will prefer to buy bread from Denis (rather than DWX-1300.0.4) because he is a friend and his daughter goes to school with their children, and they can take advantage of their purchase to invite him to join the parents' association.
In this context, the market is subjugated to the community, not the other way round, and this is precisely what some of our concerns about AI reveal: in many cases, we suffer from our lives being subjugated to capital, not because the idea of capitalization is alien to us - it is not - but because it alone is not what determines the meaning of our lives.
In this connection, I would like to delve a little deeper into the subject of intellectual and artistic production. This is one of the sectors most affected by the spread of AI. Whether it is journalism, copywriting, illustration, music composition or anything else, AI is making its way everywhere and, in some cases, landing contracts in the place of certain professionals, mainly for budget and turnaround time reasons.
This is particularly obvious in the fields of illustration and copywriting, and if a sharp eye can still spot that an image or text has been generated by an algorithm and not by a human, it is obvious that this is not going to last, and that we should soon reach situations where we can no longer tell the difference. This is often the case for the unwary.
What's missing here, however, are the other human elements involved in the action of artistic and intellectual creation. I don't pretend to know them all, but here are two main ones that seem absolutely essential.
The first is the very act of creation. Indeed, the use of AI focuses on the finished product and what it can be used for. If I write an article with this tech, what interests me is having the article available in a few seconds as a finished product, more or less.
The creation process is almost non-existent, because the only thing remotely human about it is giving instructions to the algorithms.
Yet, the creation process is not just a constraint or a suffering, it can also be a real pleasure, a leisure, a meditation, a sublimation in some cases.
Whether I write a novel, devoting days to the creation of tailor-made characters inspired by my experiences and concerns and beliefs, or ask an AI to do it for me in a matter of seconds, it won't have the same flavor or function for me
The second element, which also seems very important to me, is the sense of intimacy we can have with a creation, precisely because it is human-made.
When I watch a film, read a book or an article, or look at a painting, the experience I have is not only due to what the product as such arouses in me, but also to the fact that it has been created by a person, a human being towards whom I can feel or sense a form of connivance, perhaps similar questioning or experiences, worries, struggles, commitments, regrets and many other things.
This intimacy experienced in contact with a work is an integral part of the way we experience it.
When I look at how Albert Rigolot painted the hayfields of a 19th-century French countryside, his past existence as human is fundamental to my appreciation, as I also see in it a testimony to an era and a sensitivity:
These two forms of experience - that of creating a work and that of appreciating it - are largely degraded when artificial intelligence is involved. The only real way to restore these experiences in such a context would be to develop artificial intelligences similar to humans, which would no longer be just intelligences, but complex, sensitive forms of life towards which we could feel a form of intimacy and connivance. This is what I am going to discuss now.
"AI will replace and surpass humans."
The question of artificial intelligence also raises the question of our humanity, and all the more so as this technology is mostly designed to imitate humans. We can therefore legitimately wonder whether the very meaning of our existence as humans can be called into question by our own creations.
On this subject, it is common to hear "Yes, but we are still a long way from what a human can do." For me, this answer is not sufficient, because it can easily be retorted by "For how long?". In fact, the scope for perfecting algorithms is potentially infinite and it is therefore expected that, given the right material resources, we could one day achieve human precision, or even superhuman precision, as is already the case in industry and research laboratories. We can also cite the case of chess algorithms, which beat the world's best players without too much difficulty.
However, I would like to recontextualize what an algorithm actually does. First of all, it is produced by humans, and is therefore irremediably limited to human potential. It can, and does, produce things that humans have never produced, but these things are always based on elements provided by humans: mathematical formulas of what we manage to conceptualize with them.
The question here is: is our creativity merely a permanent recombination of elements we already know, or does it have something extra?
In the first case, we have to admit that if humans can produce an algorithm that mimics humans, then the latter can create anything that humans can potentially create.
I feel, however, that this equation is missing a very important element: mathematical limits. Indeed, contrary to what some people versed in the field think, it is not possible to express everything mathematically, just as it is not possible to express everything in words. In both cases, we are dealing with a constructed language whose role is to try and symbolize reality as best we can. However, this temptation to express things in the right way does not come about ex nihilo; it stems from an intuitive and shared feeling that this is one of the most satisfying ways of expressing things. That said, human beings sometimes disagree with each other about it, giving rise to different cultures and intercultural compromises.
On the other hand, from a metaphysical point of view, trying to explain mathematics by mathematics or language by language is a tautology by itself and always leads to an aporia that comes from the fact that any expression requires a reduction, since our thinking can only imagine finite elements and therefore only portions of reality. We can neither picture nor express reality perfectly. It is ontologically impossible.
Moreover - and this is a phenomenon of fundamental importance to our subject - humans possess the capacity to change according to their experiences in the cosmos. This change is enabled by our interactions with our environment, and takes place at levels that are sometimes incredibly subtle.
In my opinion, the most revealing example of this is the way mankind has named colors throughout history. One would think that we have always approached colors in the same way in all times and places, but that is not true: there is a fairly precise order in which we have gradually named the colors we observe today (red being the first we named and blue the latest).
This phenomenon, observable in texts, may reveal that our eyes don't see the same spectrum of light as they did ten thousand years ago, and that there are things we see today that our ancestors did not, and vice-versa. We might even think that, in time, we could observe new frequencies of electromagnetic waves as a result of the continuous, artificial emissions we inflict to ourselves. This ability would then perhaps appear as a survival mechanism to avoid harmful EMF and get closer to healthy ones. Who knows?
This capacity for adaptation would tend to prove the second idea - that we don't just recombine the known, but enrich ourselves with external contributions - right. We can, however, interweave the two ideas at once by seeing ourselves as part of a wider universe, and that it is not just the human who creates by combining what he already knows, but also the universe that recombines the human from what it is beforehand as an entity transcending all forms of measurement.
The degree of sophistication of the natural mechanisms I have just presented raises the question of whether we are capable of mimicking them to perfection, and also of the value of doing so from a certain point of imitation. For my part, I leave the door open to the idea that certain humanities might gradually move towards the ability to imitate nature to perfection. This, it seems to me, would be the ultimate achievement, for then, in a way, we would have become God. However, unlike many people who debate this subject, I don't see this possibility as a challenge or a rebellion against some demiurge, I simply consider that the levels of consciousness and power necessary for such a feat necessarily imply total fusion with the creative principle and therefore a disappearance of the state of humanity, individuality and even manifest existence. This is metaphysics, but I wanted to make it clear because it is a discussion that often arises when we broach the subject of AI, and I personally find the metaphysical prerequisites sorely lacking for such discussions to make any sense at all.
On this point, the intelligence produced would be no more artificial than the intelligence already created in the cosmos, and that would not be very surprising, since we are as much a part of the cosmos as it is of us.
"Is all this reasonable?
As I was just saying, it is debatable whether there is any point in imitating nature to such an extent once a certain level of sophistication has been reached. We might even ask whether it is worth imitating nature at all.
Ultimately, we can ask ourselves what we are trying to achieve through the increasingly complex development of tools that imitate the living world. First of all, there is the scientific curiosity that has driven human beings for ages, but which has not always been decorrelated from philosophy and morality, unlike today. Then there is the capitalist obsession for new solutions to maximize profit while minimizing investment.
There is also, I believe, the fascination of having the power to create, which brings us back to ancient demiurgic hopes that are an integral part of humanity's existential questioning. However, these hopes are sometimes unhealthy, in my opinion and especially in our time, because they seem to me to ultimately seek to exercise a form of omnipotence that would suffer no responsibility, an equation in fact insoluble which amounts to wanting to be the cause of no consequence. A one-way ticket to nothingness.
And yet, if we are to reach the divine in our creations, I believe it can only be done by accessing a form of mastery over reality that will inevitably remind us of our responsibility and the fact that our actions do have consequences.
Because, yes, we already know how to create incredible intelligences, that is called having children, and we already know how to paint canvases, that is called becoming an artist, and we already know how to make smart tools, that is called becoming a craftsman. But wait, there is more, because these three talents sometimes coexist within the same person! And all this, with relatively low energy consumption: three meals a day and a few rest periods.
But it takes time, it is not always easy, it is not always straightforward, and we have to face failure, doubt and suffering – it is all part of the game.
By creating artificial intelligence, we certainly avoid some of these problems, but at what cost? Because, and this is something rarely said, all these ultra-sophisticated artificial creations are also extremely energy-hungry, overall, and for what they ultimately provide in terms of value: they don’t know how to replicate themselves, nor how to repair themselves completely, nor how to collaborate spontaneously, nor how to change their methods when necessary, nor how to feel the beauty of existence, nor how to develop spirituality or a value system, nor do they have compassion, hope, desire or madness.
In conclusion, I would say that the only value artificial intelligences have is in the experience humans live in creating and using them, and I personally believe they can never surpass the experience of the cosmos, or the experience of God, for the ones comfortable with that term.
That is why, personally, I don't spend much time on them. In fact, I only use sophisticated algorithms of this kind in two cases: for translation from one language to another, and to find information that the bubble in which search engines enclose me prevents me from reaching.
But the fact remains that, in the first case, I am always re-phrasing afterwards to try and express exactly what I want (and what the algorithms will never know for sure) and that, in the second case, I am always more successful by talking to people who have the skills and knowledge I lack to do or learn something.
In the end, if I already lived in the kind of environment I aspire to, I would not even think about these inventions because I would not need them at all.
There are already more fabulous things in our cosmos than I will ever be able to experience in this lifetime, so the degraded experiences proposed by artificial intelligences seem to me to have the blandness of a tomato produced by the chemical industry, even though I already have plenty of healthy, delicious tomatoes in my garden.
Another excellent piece of writing. Thank you. 🙏 I really enjoyed the thought process and the conclusion made sense to me on an intuitive level.