Paroled from the Paywall
What Minds What We Matter?  

3,578 words

Part 1

Jacques Luzi
Ce que l’Intelligence artificielle ne peut pas faire
Èditions La Lenteur, 2024

For both pragmatic and mystical reasons, industrial society worships technology, and it is the adoration of technology which stymies any democratic debate or questioning regarding the legitimacy of the technocratic class in power. The dependency on machines and the devotional association which many ordinary people have with machines stands in the way of any effective critique of that class’s project to transform natural gifts into artificial ones.

That extract (p.25), taken from the preamble to his short essay on cultural intelligence, What Artificial Intelligence Can’t Do, is typical of the timbre of Jaques Luzi’s work and points to what it is really about. It is a rallying cry against Artificial Intelligence, not a clarification of what AI can or can’t do. Therefore readers expecting to be provided with a sober informed account based on technical data of the capabilities and potential or possible limits to the powers of A.I. will be disappointed. What Artificial Intelligence Can’t Do is not a continuation of Hubert Drefus’s analytical critique published in 1972, What Computers Can’t Do, albeit Luzi’s title is probably a nod in Drefus’s direction. Those who who are looking for such a continuation will be better served elsewhere, by literature on the subject online or in print. What Artificial Intelligence Can’t Do is not a scientific analysis, it is a polemical, impassioned and unashamedly partisan critique of AI.

Although arguably misleading, the title of his book is nevertheless not entirely extraneous to what Jacques Luzi is attempting. His critique of Artificial Intelligence is based on his understanding of it as something inhuman and ultimately anti-human and the logical conclusion to be drawn from such a viewpoint is that what AI “can’t do” is be Homo sapiens but what it can do is destroy humanity.

By claiming that AI is an assault on the soul and by refuting the claim that AI can redeem or even ultimately benefit humanity, Jacques Luzi is sounding a tocsin. AI, he warns, will not benefit mankind in any sense, although its many advocates like to claim as much for it; it cannot save the planet from environmental deterioration or ecological collapse (the reverse, according to Luzi). This is the direction of the essay, so the title may indeed seem misleading to many. More appropriate titles might be Artificial Intelligence and the War against Nature, or The Chimera of Technical Progress or even Artificial Intelligence and the Denial of Redemption.

Luzi’s language is religious even though he does not talk about God. With its argument that AI is the logical consequence of an obsession with growth and progress, What Artificial Intelligence Can’t Do is also, for better or for worse, reactionary, for it is written in reaction to a global consensus. That consensus states that economic progress is by definition desirable. Inseparable from faith in AI, so Luzi argues, is the assumption that progress must be the aim of all social and political endeavor, therefore that all social problems and challenges must be solved by enhancing technical speed and efficiency. This is achieved by advancing scientific knowledge and applying such knowledge to confront every challenge which mankind may face. For its proponents, AI is the rational acme of necessary and ever increasing technical efficiency.

It is surely true that underlying all political and economic debate lies a rarely challenged assumption, to the effect that Growth and Progress, (often referred to in reverential terms as though they were abstract forces of Goodness and written as proper nouns) are necessary and beneficial by definition. It is that assumption which Luzi’s critique of AI seeks to rebut.

What Artificial Intelligence Can’t Do is not focused on AI in isolation. It links advocacy of AI to a blind veneration of Growth and Progress. By questioning the benefits of Artificial Intelligence, this essay leads the reader to wonder whether growth and development are intrinsically desirable, and if they are, why they are. Luzi argues that AI is a weapon of war against all that is natural, an assault on the natural world, a quest to destroy natural processes and replace them with mechanical ones.

Luzi’s work focuses on a few key fields of AI application, whereby it is unclear to this reviewer whether he highlights them as examples of AI, or points to them because they are so negative as to undermine belief that AI can be beneficial in any other respect. The applications of AI considered are its current and potential application in war, its role in destroying ecological balance and the role it plays in what the author perceives as the fraud of modern “green” policies, and finally, the increased opportunities which AI provides to the state to place its citizens under permanent and comprehensive surveillance and control and drastically restrict their freedoms. The argument is also that AI makes people weaker and more passive, because it enhances automatic problem solving efficiency at the expense of encouraging individual initiatives to solve problems.

Contemplation of Artificial Intelligence leads to a consideration of evolution and eschatology. In what direction does anyone believe that Homo sapiens could or should evolve? Should our species even continue to evolve at all? After all, AI is indisputably an evolutionary development, (everyone can agree on that) involving the evolution of skills, the self-adjustment of machines to given data, and the ability of machines to react rationally to new circumstances, just as hominids adapted in the course of their evolutionary history to become what they are today. The aim of the promoters of Artificial Intelligence is to develop a non-human and inorganic entity, “intelligent” in the sense that it can make autonomous decisions, replacing the idiosyncratic, faulty and slower calculations and decisions made by human beings, “artificial” in the sense that it is driven by inorganic mechanisms not found in nature.

Can and should the application of decisions made by human beings be influenced, let alone driven, by virtue of intelligent interventions to change natural developments, and if so, how far should those interventions be taken? To the point that the decisions themselves are made outside and even in contradiction to decisions made by human beings? Will Artificial Intelligence ever be able to affect the evolutionary course of Homo sapiens and if it ever could, would or should we welcome that?

It should be borne in mind that many interventions to alter human biological processes, including reproductive ones, are carried out all the time without the use of Artificial Intelligence, for example the contraceptive pill and medication in general. There have also been “artificial” but not AI interventions to alter the natural course of the biological evolution of the entire species, a significant example being the reversal of the evolutionary bifurcation of human races, that is to say the subspecies of Homo sapiens, from a course of species diversification. Permanent isolation of the three sub-races of mankind would have led in time to the emergence of new species had it not been for the “artificial” intervention of recent migration, colonization and integration policies and policies based on the profoundly anti-scientific doctrine that subspecies/races of human beings simply do not exist. Nevertheless, Artificial Intelligence hugely increases the potential for such interventions.

Will AI always be the servant in such interventions, or one day the master? Is mankind, as Nietzsche foretold, something fated to be overcome by a superior being, an Übermensch? Will there be a future time when entities that have evolved out of and beyond Homo sapiens regard our clownish and clumsy efforts in life with the same condensation as mankind today regards the behavior of monkeys?

These profound but not technically complex questions arise from a reading of Luzi’s work. Beyond AI per se, it is mankind’s blind faith in the benefits of departing from nature which is the subject of Luzi’s invective. He believes that mankind nurses within himself an impulse to destroy nature and from that a will to self-destruct. An occult contempt of our species towards nature in general is taken by AI from a hitherto more or less unconscious level of contempt to a blatant and ultimate one, ultimate, because Luzi believes that AI is destined to supplant nature and along with nature, the species which created it. For Luzi, AI is the vehicle which severs the bonds which link Homo sapiens to the natural world, which forces the species to take the quantum leap from natural man to mechanical “man”.

The chapter titles of What Artificial Intelligence Can’t Do serve as bullet points to Luzi’s essay and provide an overview of it:

• A Martian writes a Letter to Elon Musk
• Artificial Intelligence: from a Conception of the Machine to the Dream of Universal Control
• The first use of Artificial Intelligence is to make War
• The War on Nature augmented by Artificial Intelligence
• The Intelligent Control of Populations by the State
• For the Human and the Natural: against Machines
• Interview with a Creator of Artificial Intelligence

Jacques Luzi’s Martian seems well read in the literature of planet Earth, for his letter to Elon Musk is preceded by a quotation from the German poet Hölderlin. It is the opening paragraph of the book:

Were it possible, we would leave the empire of the Sun to take the leap beyond the orbit of our wandering star. But alas, there is no possible homeland for Man other than the one which harbors his heart.

These words and the introductory letter call to mind the possibilities of space exploration, although that subject does not feature in Luzi’s treatise. One may safely assume that Luzi is less than awed by the idea of exploring other worlds. Those who reject AI in principle must needs reject space technology, for it is becoming increasingly AI dependent. If AI promises to release Homo sapiens from the bonds of nature, it also promises to free Homo sapiens from confinement to planet Earth.

Luzi’s Martian writing to Elon Musk notes that according to his biographer, Walter Isaacson, Musk’s space ambitions began when he read The Hitchhiker’s Guide to the Galaxy. There, the reader may recall, technology with super human capacity is harnessed to save the world from the destructive force of… technology! The irony is not lost on Luzi, (but it seems entirely lost on most advocates of AI) who points out that AI is often justified as a technological solution to a problem or challenge that was created by technology in the first place.

Luzi’s Martian alleges that Musk is autistic, humorless and insensitive, and states that the founder of Tesla is so lacking in deep perception that he misunderstood Mike Judge’s film Idiocracy. In Idiocracy a man awakes from a deep sleep of many years, Rip van Winckle fashion, to a world in which successive generations of low IQ folk breeding more rapidly than their intellectual superiors has resulted in a population which is borderline imbecilic. Homo sapiens, no longer sapiens, is on the road to self-destruction by reason of its own dysgenically induced stupidity. For our Martian, and for Jacques Luzi, who is clearly speaking through him, Idiocracy is a “satire on eugenicism” and Musk reveals his anti-humanism (for Luzi anti- humanism is a hallmark of those who advocate AI) by accepting the eugenicist message of the film at its face value, without grasping that the film is satire. Luzi may have had in mind the fact that Musk was once diagnosed with Asperger’s Syndrome, which belongs to the autism spectrum disorder, and a characteristic of autism is difficulty in recognizing irony. On the other hand, it is not clear to this reviewer that Idiocracy was indeed intended as a satire on eugenicism.

Be that as it may, Luzi’s Martian reminds the reader, even if doing so with a debatable point, of the attraction which efficient and “soulless” measures hold for those like Elon Musk, who are enthusiastic promoters of AI “solutions”. That critique strikes home: AI solutions are by their very nature devoid of empathy, of feeling, of human engagement. Complex social or natural problems can be solved with ruthless efficiency by AI, including health, military challenges, law and order and more. However, neither humor nor mercy nor mitigating circumstances are qualities likely to feature in an AI directed program. Luzi implies, and this reviewer has no reason to doubt it, that advocates of AI tend to be enthusiastic advocates of assisted suicide protocols, abortion on demand policies, the eugenicist selection of optimal human embryos and other measures which may appeal to the mind but surely never to the heart. Such efficient solutions are, like AI itself, gaining favor in many parts of the world, notably in Canada, where doctors are now empowered, on the strength of what appear to be outright eugenicist “MAID” (Medical Assistance In Dying) protocols to advise financial “no-hopers” such as despairing teenagers, people in great pain and other social “losers”, on medically assisted suicide, and if consent is given, doctors are authorized to administer a lethal injection to them. The integration of AI into euthanasia programs would help to remove any element of personal responsibility from those working the program and would offer a tempting way of disengagement from the problematic human aspects of it.

Luzi believes that a focus on efficient solutions to social problems, using AI as a tool to solve them, can and indeed will encourage and accelerate such programs and protocols, for example the selection of desirable embryos before birth (and disposal and marketing of the undesirable ones). In addition, he believes that the expansion of the role of AI will be accompanied by a depletion of human cognitive skills due to the increased reliance on mechanical AI “solutions”. That sounds plausible: just as bird species which do not use their wings, like the penguin, eventually have what were once wings atrophying to become useless appendages; in the same way, when humans cease to exercise their cognitive skills, for example when AI doing cognitive tasks for them, human IQ’s decline. Luzi also argues that the expansion of AI greatly assists the state in expanding its power. As human social interaction becomes more infrequent thanks to AI, and less efficient by comparison with the AI solution, personal freedoms are eroded by the power intrinsic in technical innovations. Efficiency and even social legitimacy, the right to life itself, will be determined by a non-human entity in the service of an increasingly dictatorial state. Luzi’s is a pessimistic vision, but recent developments in the world lend it credence.

On the other hand, space science could not advance without AI. In recent years mankind’s knowledge of the cosmos and of celestial bodies in this solar system and beyond it has been increasing dramatically. AI has played a major role in expanding that knowledge and thereby in breaking down “provincial” preconceptions of the uniqueness of Earth and the impossibility of leaving it. Up until the last years of the twentieth century, most people, encouraged by many scientists with views echoed in most published literature on the subject, regarded it as unlikely that planets existed outside the solar system. Now in the wake of discoveries of numerous exoplanets, “the solar system” is often referred to as “our solar system”. Until recently it was widely doubted that elements necessary to life, notably hydrogen and oxygen, would be found on moons or other planets or that moons and planets even existed beyond our solar system. However, since the 1990’s, when the first sighting of an exoplanet was confirmed, over four thousand exoplanets have been officially recognized, with every chance that thousands more are awaiting discovery in the near future, while it is now stated that H2O exists on our moon and very probably on other moons and planets too, both within our solar system and beyond it.

The notion of a vast and even expanding cosmos, populated with hundreds of thousands of planets, allocates the Earth to a humbler place. If AI is “dangerous”, as Luzi maintains, it is also true that it leads the way to great adventure and great discovery beyond planet Earth and a less arrogant conception of the Earth’s uniqueness. Luzi is silent about this. However, his work may lead a reader to perceive a dilemma. If AI offers practical benefits in some fields (Luzi says nothing about AI advances in medicine for example) and opens the way to space exploration, what of its evolutionary threat and the potential enhancement of an all seeing all powerful state or omnipotent world government?

If it is perilous to accept AI in its entirety, is it feasible to reject it in its entirety? Presumably Luzi himself used a computer to write What Artificial Intelligence Can’t Do and probably he sent his copy to his publishers by email. One comment by Luzi’s Martian in his letter to Musk indicates that the author’s rejection of modern science is sweeping: “the idiots in the film water their crops with soda pop, just as the idiots of the present are spraying their crops with pesticides.” (p. 12).

Watering with soda pop and spraying with pesticide surely do not reflect the same levels of idiocy, even if one concedes that in the long term, spraying the land with pesticide may be deleterious. Watering crops with pop will lead them to perish, spraying crops with pesticide may protect a harvest. Does Luzi regard all interventions in nature as foolish and ultimately destructive? If so, that could be seen as a Luddite extreme in contrast to the extreme of venerating Progress as a never-to-be-criticized abstract, the target of Luzi’s invective in his treatise. One might say that Luzi’s is an extreme Rousseau like “back to nature” reaction to an extreme Ayn Rand like position which affirms that technical Progress and rationalism are intrinsically good.

Again, the essay raises a question or dilemma without addressing or even acknowledging it, namely how much of a risk should be accepted in technical innovation weighted against a perceived benefit? Is the evolution of science always to be welcomed and does it necessarily augment the sum total of human happiness? A consideration of simple examples highlights a conundrum. Most people will agree that the invention of the electric light constitutes progress in terms of both domestic and industrial efficiency, one which in respect of a Benthamite utilitarian calculation raises the overall level of human felicity. Nevertheless, not many people favor electric over natural light in hours of darkness all the time and in every situation, and certainly not compared to sunlight. Even whilst agreeing that electric light is more efficient and practical than candle or oil-lamp (and less dangerous), many will opt for natural old fashioned alternatives in certain situations.

The romantically inclined still prefer candlelight for an intimate evening dinner for two. Less efficient though it be, candlelight is a natural light and it is natural light which people find more restful, more soulful, more romantic and even more healthy, and it is true that people find more comfort in what is natural and familiar to them, in what is organic, compared to the mechanical, the artificial, the inorganic. “It’s not the same thing.” is a frequent instinctive explanation of a reluctance to accept the artificial as an alternative to the natural, even when the artificial is evidently more practical or less costly, as it often is. What is romantic is closely linked to what is natural. The mechanical is instinctively apprehended as unnatural, which indeed it is. And who can disagree with Luzi when he argues that artificial improvements tend to take humans away from “where the heart lies” ? Another example: most people if asked would probably say that the replacement of road by rail and then rail by highway constituted technical progress, one which increased the sum total of human happiness. However, few people would want highways built through every small town, and still fewer would welcome highways leading up up to their front door, however efficient and progressive that might be.

Luzi strikes this reviewer as going to extremes with his dislike of all science, but those who dogmatically dismiss skepticism about the benefits of technical change, who decry skeptics as “out of touch”, “unrealistic” or “standing in the way of progress” are extremists themselves, and those who venerate Growth and Progress wield incomparably more influence in global decision making than those who yearn for a return to the Garden of Eden.

An important fact to bear in mind is that by nature of its evolutionary success and support, and by virtue of the power of the rational and efficient, the AI solution acquires its own momentum and tends intrinsically to gender further “improvements” and to extend its capacity to destroy what is traditional, customary and time honored. Holding on to a way of doing something out of sentiment or loyalty gives way inexorably to the rationality of introducing “efficient” AI solutions. AI is constantly and rapidly evolving and it is “in the nature” of an evolutionary force that it is contemptuous of sentimental attachment, that it replaces less efficient competitors by a process of “natural” (!) selection without regard for any factor other than efficiency levels. The only question which concerns the AI advocate is “how well does this work?”

One can argue that it is the impassioned reaction of the heart, as shown by Luzi in What Artificial Intelligence Can’t Do which is necessary to counter the all too reasonable, rational blandishments of an omnipotent future world ruled by AI.