The title of Aakash Singh Rathore’s book, Will AI Murder Us All?, refers to the Fermi paradox: if the universe is billions of years old and teems with stars and planets, where have they all gone? Statistically, extraterrestrial civilizations should be common, yet we have observed no evidence of their existence: no extraterrestrial signals, no artifacts, no visits. The proposed solutions range from “We are alone” (the rare-Earth hypothesis) to “They are hiding” (the zoo hypothesis) or “They have destroyed themselves” (the coming Great Filters). It is a paradox that fuels existential dread and science-fiction dreams, reminding us how vast and silent the universe appears.
How then are civilizations filtered? The reason civilizations do not propagate across the stars is that, just at the moment they develop technologies powerful enough to do so, they self-destruct in some way — they are thus filtered — even if only a tiny minority of inhabitants tends toward self-destruction or is prone to take big risks, sooner or later one of them will seize the opportunity… Aren’t we precisely at such a moment today? We have created an artificial intelligence that threatens to control us completely, erasing what makes us human. We have exercised domination over our natural environment to such an extent that we render the Earth uninhabitable for ourselves. We have developed weapons of mass destruction that, if deployed, would annihilate not only our enemies but us all.
The book by Singh centers around this single question: how, if at all possible, can we — humanity — avoid being “filtered”? It describes our situation with merciless candor, rejecting all easy humanist excuses: “A future AI, if it surpasses us more generally, could relegue human cognition to oblivion — not out of malice (though that is possible if one considers a broad techno-panic like Isaac Asimov’s classic book, and later the film with Will Smith, I, Robot), but precisely by being better at being less human.” The gap separating AI from the human mind is not merely quantitative, but qualitative: an AI agent does not merely learn data and reason much faster than a human mind; more importantly, the human mind is embodied, exposed to pain and vulnerability, subject to doubt and fluctuations, trapped in an intersubjective network with others, which are no less opaque to themselves than the human mind that interacts with them, a contrast to an AI agent that is a machine operating smoothly, simply performing its task without compassion or pain. But can such an incorporeal, flawless digital machine (which far surpasses our intelligence) possess a form of self-consciousness or consciousness of any kind? In a highly problematic way (for me, at least), Singh dismisses this question as irrelevant:
“Machines reason, learn, create, communicate and already solve novel problems better than many humans in specific domains. If they perform the function of thinking, perhaps they do think, with or without consciousness. Without feathers.”
How then will AI agents with intelligence far beyond human tackle humanity, more precisely, human morality, which characterizes us? The obvious first thought here is that AI agents will have no morality (and they already have none): notions such as guilt and responsibility are foreign to them; they simply follow the moral rules their programmer has implanted…
The next option is that AI agents will operate like the Übermensch (Overman) of Nietzsche — Singh is right to emphasize that the Overman does not merely develop our moral values to elevate them to a higher level; on the contrary, “the Overman creates his own new values and is ready to do exactly the opposite of what tradition dictates, if that enables the discovery of new possibilities. Seen from this angle, the super-intelligent AI might not so much constitute a danger to morality as a possible and welcome transvaluation of it: a phenomenon that redefines what it means to be excellent, creative, or virtuous. The modern discourse on AI — from Elon Musk’s articulation of artificially “divine” minds to Peter Thiel’s rejection of democratic constraints on technological power — perfectly captures this Nietzschean sensibility: the future belongs not to those who preserve inherited morality, but to those who risk everything to transcend it.”
Singh convincingly rejects this Nietzschian reading, advancing two counter-arguments. First, since AI agents will be far smarter than humans, they could obey their own morality — a morality that we, as humans, will not be able to identify or even understand:
“An AI might reason about suffering, fairness, or the future value of humanity in a way that is internally coherent, yet completely alien to human moral intuition. In such a case, the gap would not be one of intelligence but of moral intelligibility.
At that point of moral singularity, humans would confront a paradox: we would be, directly or indirectly, at the mercy of a system whose ethical judgments we cannot evaluate, but to which we cannot resist. In that case, super-intelligence would not guarantee super-morality, but could instead generate total moral opacity.”
The second reason offered by Singh is even more interesting and pertinent: Thiel and Musk may regard themselves as overmen—creative individuals who stand apart from the herd—but what they are doing in fact is simply regulating, or absorbing, our herd-like and conformist behavior: “the danger or lure of AI does not lie necessarily in its supplanting our existing, outdated morality, but in its reinforcement of our tendencies toward conformity and herd behavior.”
Thiel’s world is not that of domineering, humanoid overmen, but of “ordinary corporate androids animated by a hive mind, a robotic workforce enslaved to corporate bottom lines…” This world is as far as possible from what Nietzsche had in mind: a new world of exceptional individuals willing to lead a passionate and intense life, full of risk, outside any universal moral norm.
So, where does Singh see a glimmer of hope (if not a solution)? His baseline postulate is that simply fighting for new ethical values that would motivate us to act for global justice will not suffice: whatever its content, such an approach, by its very form, reproduces domination and exclusion. As with Buddha, our starting point should not be universal principles but the concrete experience of suffering, humiliation and exploitation that we observe around us. If suffering is part of the human condition as mortal, vulnerable beings, trapped in fragile bodies and rooted in a world of unique life, that vulnerability is today exacerbated by three factors: the tradition of patriarchy and colonialism that survives in a disguised form in our modernity; European modernity, which presents itself as bearing equality and justice but uses this ideological framework as a tool that only reinforces exploitation, oppression of women, and new forms of colonialism; and, above all, the explosive growth of artificial intelligence, which controls us precisely at the moment we perceive ourselves as free. To see this, it is not enough to proceed by the usual mode of ideological critique and to expose the inherent inconsistencies and exclusions of modern universalism; what we must focus on is how these three factors structure our daily life, how we experience their effects in our suffering and our daily disorientation. About AI: what does it mean for an ordinary citizen to live in a world governed by AI? How does AI enable those who hold power to know themselves better than they know themselves, how does it predetermine their desires and their choices?
Singh is also entirely right to locate the threat posed by AI not in the technology as such, but in the social relations within which AI-enabled machines are produced: it is the capitalist universe that prevents us from escaping the logic of the prisoner’s dilemma. Although we rationally know that solidarity and cooperation are better for all participants, the emphasis of capitalism on profit privileges suspicion toward others — it perceives others as potential threats who will use our trust as a pretext to attack us even more brutally. The fundamental perversion of capitalism lies in the fact that it is not even truly self-interested: what matters is not only that I win, but that my opponent loses, so I would rather lose a little (if my opponent loses more) than win a little (if my opponent loses more). We approach here the theme of envy and resentment, which cannot be dismissed as mere subjective aspects of an alienating social order — they constitute a constitutive element of human desire itself, clearly perceptible even in leftist projects that advocate equality and solidarity, which often rest on the premise that I am willing to suffer only if others suffer as well…
How then to escape this vicious circle? In addition to a broad sympathy for Buddhism, Singh here cites two authors, Donna Haraway and Carol Gilligan. Haraway locates the emancipatory potential of AI in the very feature that most perceive as a mortal threat to our humanity: we should abandon all anti-technological nostalgia for older modes of “authentic” existence and embrace with joy that our lives, even in our most intimate feelings, are inextricably entangled with technology. In the most literal sense, we increasingly function like cyborgs, integrating technological components into our bodies without which our social life could not survive, and AI is just another, far more powerful technological complement, without which our entire social life would collapse instantly. By their very existence, cyborgs break with the logic of parental succession — they are an “illegitimate offspring, born not of love or emancipation, but from domination: militarism, patriarchal capitalism, state socialism, war technologies, bureaucratic control, industrial discipline and systems designed to manage bodies and populations. The cyborg is a by-product of violence and exploitation.” Such a practice of reappropriation (or exaptation, to use Stephen Jay Gould’s term) uses technologies generated by oppressive systems against those very systems:
“This is a radical reorientation of how we conceive the technological heritage. Where traditional political theory — and one could even include Gandhi and his demonization of modern technology — worries about corruption and degradation (the way good intentions are corrupted by bad systems), the cyborg transforms the instability of its original systems into a weapon, thereby opening the way to new epic possibilities.”
It should be noted that the entire narrative of the origins of Islam rests on an illegitimate lineage as well. A crucial distinction between Judaism (and its Christian continuation) and Islam lies in the fact that, as seen in the case of Abraham’s two sons, Judaism chooses Abraham as the symbolic father, the paternal symbolic authority, of the official line, thereby rejecting the second wife. Islam, by contrast, opts for the line of Hagar, the Egyptian slave who bore Abraham his first son, Ishmael (the ancestor of all Arabs, according to the myth). Like Sarah, Abraham’s legitimate wife, could not become pregnant, she instructs her husband to lie with Hagar. But after Sarah bears a son, Abraham sends Hagar with Ishmael into the desert where they are supposed to die of thirst. God intervenes directly and saves them both. Although Abraham and Ishmael are mentioned dozens of times in the Quran, Hagar is not mentioned, effaced from the official story. Yet she continues to haunt Islam, her traces surviving in rituals, such as the obligation for pilgrims going to Mecca to run six times between the two hills of Safa and Marwah, a kind of neurotic repetition of Hagar’s desperate search for water for her son in the desert. Muslims thus perceive themselves as the descendants of an illegitimate child: their founding figure is an orphan (and the same is true for Muhammad himself). The (sadly) famous umma, the “community of believers” in Islam, is a community “coming from nowhere,” in the desert genealogical, like a revolutionary and egalitarian brotherhood of orphans — not surprising that Islam experiences such success when young men find themselves deprived of the traditional family security network.
Returning to Haraway, she soon abandoned this cyborg dream: the gap separating the human part from the AI part of a cyborg is too deep, such that it is hard to imagine a community of cyborgs programming their AI part to follow the logic of benevolence and solidarity. The dream of the illegitimate child revolutionizing the established order remains politically important: it allows us to discern the line of political history that began with Han Fei (the founder of Chinese Legalism) and passes through Machiavelli and Saint-Just to Lenin, yet another genuine “illegitimate child” of Marx (in the split between Bolsheviks and Mensheviks, Marx’s legitimate offspring was certainly Ioλ Martov, the leader of the Mensheviks). Lenin thus sits in a long line of political figures who began with the Chinese legalists emerging from the Confucian crisis. When, between the 5th and 3rd centuries BCE, China experienced the period of the Warring States, the Confucians perceived in the treachery of old traditions and customs the ultimate cause of this slow but persistent decline. What troubled Confucius most was his sense that the political institutions of his time had collapsed. He attributed this collapse to the fact that both those who wielded power and those who occupied subordinate positions did so by claiming titles they did not deserve.
Asked about the principles of good governance, Confucius would have answered: “A good government is that the ruler is a ruler, the minister a minister, the father a father, and the son a son.” In Europe, we call this a corporative view: society is like a body in which each individual must stay in his place and play his particular role. That is the exact opposite of democracy: in democracy, no one is forced into a particular place; everyone has the right to participate in public affairs and to have a say in deliberations about the direction of our society. It is therefore not surprising that Confucius’ description of the disorder he observes in the surrounding society — “The rulers do not govern, and the subjects do not serve” — corresponds well to a democratic society in which the subjects, united, govern and the nominal rulers serve them.
What the Legalists did was to abandon the very foundations of such a perception of the situation: for the Confucians, the country was in chaos because the old traditions were not respected, and states like Qin, with their centralized military organization that despises old customs, were perceived as the very embodiment of what is wrong. Yet, unlike his master Xunzi who regarded nations like Qin as a threat to peace, Han Fei “proposed the unthinkable, namely that perhaps Qin’s way of governing was not an anomaly to be corrected, but a practice to be emulated.” The solution lay in what seemed to be the problem: the real cause of trouble was not the abandonment of old traditions, but these very traditions themselves, which daily demonstrated their inability to serve as guiding principles for social life — as Hegel phrased it in the “Preface” to his Phenomenology of Spirit, the standard by which we evaluate a situation and determine that it is problematic is itself part of the problem and must be abandoned. Han Fei applied the same logic to the fact that (the majority of) men are bad by nature, not inclined to act for the common good: instead of lamenting it, he saw in human depravity a challenge for the establishment of laws severe enough to control this nature and to use it for the benefit of the State.
One of the great accomplishments of contemporary leftist political theory (Althusser, Balibar, Negri) has been to rehabilitate Machiavelli, to rescue him from the usual “Machiavellian” reading. Since the Legalists are often presented as the original Machiavellians, one should do the same with them, extracting a radically emancipatory core from their dominant image of proto-“totalitarians.” The great intuition of the Legalists was to perceive the wound (inflicted on the social body), the disintegration of old habits, as a chance for a new order — though one might doubt that our “decentering” by the machine of AI could transform into a new order in which we, humans, would still be capable of reprogramming our AI partner to act with care and empathy. The Legalists advocated a new order based on the total power of a centralized state that brutally imposed its conception of a just legal system — is it not almost certain that AI will impose itself as an even stronger power whose very functioning would transcend our understanding? This is where I see the limit of Singh’s reference to Carol Gilligan, who advocates “a feminist approach to human values (and perhaps to AI) that prioritizes compassion, empathy and understanding in which individual well-being emerges from a broader social system of care, rather than masculine values such as autonomy, power and control (or, one might call it, a justice ethic).”
I find this passage problematic on four counts. First, I reject the simplistic and traditional opposition between masculine and feminine values and approaches on which the reasoning rests: I believe this opposition is not only historically limited, but also that the two poles are strictly complementary — there is no Antigone without Creon, to cite the most famous historical example. Second, even if humanity (but which humanity? We all know AI is controlled by a small circle of state and corporate devices…) somehow managed to “decapitalise” AI, to reprogram it to follow a non-capitalist path of compassion and solidarity, would that be the agent of this radical social transformation? I don’t think Gandhian-feminist propaganda could succeed in waking consciences. Third, I think Singh underestimates the tension among the three obstacles to a new solidaristic society: traditional patriarchal, racist and sexist hierarchies; capitalist modernity; and the impact of AI. Modernity has also opened the prospect of challenging those hierarchies; its values are not merely an ideology masking hierarchies and exploitation. The explosion of AI is gradually transforming global capitalism into something new that we still struggle to name (“digital feudalism”). Elon Musk is no longer a capitalist in the older sense. As for the notion of benevolence, universalizing it is very difficult: benevolence and compassion are generally limited to one’s local community, and universalizing them carries the risk of imposing our values on others. Benevolence should therefore be complemented by a more abstract notion of justice which, by its very inaccessibility, reminds us of the limits of our lived world.
This implies that universal values and notions cannot be reduced to the forced universalization of particular ideological spaces, thereby obscuring the way these spaces push their own particular interests that, in effect, exclude others. Isn’t the history of the notion of “human rights” the history of its gradual extension to those who were initially excluded: first women, then other peoples (let us recall the Haitian Revolution as a vast echo of the French Revolution), then workers…? More than compassion and benevolence, the dimension of universality implies a semantic openness that makes possible the inclusion of an emancipatory content. The exemplary philosophical illustration here is the Cartesian cogito: it is generally rejected as an empty abstraction of a disembodied thought, modeled on the ideological myth of masculine self-control; yet how many philosophers know that many educated women have given it a feminist turn? Their motto was “cogito does not have a gender,” which means that it entails an equal principle among the sexes, marking a clear break with traditional sexual ontology (as Aristotle believed the masculine form shapes the feminine matter).
The inevitable conclusion, therefore, is that AI obliges us not only to rethink what the “being-human” will become in the AI era, but also what it has been up to now — this is where the question of consciousness and subjectivity enters, and it must not be ignored.