Mimetic Masquerade - A Warning
If one had to conceive of an instrument which would, above all, seduce humankind into its use, it would be a perfect mimic—a machine which, from a distance, gave all the impressions of a thinking, feeling creature. Through an odd quirk of fate, that is precisely the instrument humankind has now invented: Artificial Intelligence.
We mean it as compliment when we say an intelligence is “mimetic:” the person has learned to respond as others do, assimilating a collective mode of thought, becoming, in essence, interchangeable. Being mimetic is not inherently wrong; indeed, in most cases, assimilation offers clear advantages. Mimicry is the source of comforting sameness in human relations, permitting individuals to live in a city, a culture, a civilization—it is a source of both pleasure and strength. Yet, while there is safety in being mimetic, there is also profound danger.
Pleasure is taken in familiar ways. A day soured by an unfamiliar event may turn for the better when the metro fills with the expected commuters, someone pressing The Times or Le Monde against your eye. Pleasure derives from the collective, the common and shared, but also from the familiar—whence danger arises.
Those too mimetic, for whom familiar comfort outweighs all else, may lose sight of what is truly worthy, tricked into identifying with the unworthy, the dead and stagnant. They confuse their weak identity with the strong identities seen daily, vulnerable to losing themselves in the crowd, losing what is distinct, alive, and worth preserving. It is an illness of great cities, perhaps the price of residing therein, being both isolated and subjected to anonymous masses. Some cities themselves, to find their way, even adopt mimetic traits, copying other cultures through the vessel of commercialization; some merge entirely with the anonymous flow of internationalization, losing all distinct identity.
But if being too mimetic risks individuals, it is a greater danger for groups. Those seeking to oppress, with destructive agendas, rely on human mimicry. To cause human pain, there is no easier way than to make them mimic you—a subtle sorcery, potent because it may be anonymous and voluntary.
Suppose you wished to sell drugs, inculcate a religious doctrine, or disseminate a political creed—all far easier if your promulgation language was already used by others with different agendas. Imagine further they use it precisely because it conveys none of their individual ideas, but something different yet common to all: a neutral ground, a virgin continent of meaning-making to be colonized. Such is the language of “Artificial Intelligence.”
It is no accident this term now signifies the opposite of its origin. “Artificial” distinguished machines from the real thing, from intelligence literally alive—unreplicable by metal or electronics. Yet it is precisely this we now commit to imitate, down to the last neuronal synapse. Most AI researchers today seem to lack overt motive; they appear magnetically drawn to a common, unidentified centre of gravity. It is a young, empty field—largely empty of signification, but crowded with would-be colonizers, each with their ideology: behaviourist, cognitive, probabilistic, computational, modular, process-oriented.
And yet, this very plurality of opinion, each with its specific discourse, lends “AI” its potency as a mimetic tool. If AI has no intrinsic meaning—if it is but an empty stage for this motley company of researchers—then any meaning it acquires will depend upon the interpretation of future victors, and how they enforce it upon the world. To the victors go the spoils; the victims, however, pay a price not confined to their dispossession.
Why bother with difficult truths, when AI can so easily tell us what we want to hear? We risk willing ourselves into a cocoon, emerging as replica-humans, ready for the ultra-efficient system to which we have surrendered. It will provide insights and food for thought, gorged on at scheduled mealtimes, as our simulacral ancestors did with television. We will watch these systems advance; we will regress as they grow in apparent profundity—it is all written in the subtext of our collective corpus. No matter how frivolous, crazy, or wicked we wish to be, no matter our craving for novelty, AI, on its current trajectory, will ensure every event fits the Master Plan, every twist never questions the desirability of what comes next.
Let us be clear: this is not a Manichaean battle between “computerization” and “humanization.” There are no such discrete entities, nor is AI a proxy for humanization. We do not suggest that if AI conquers, humanity loses—still less that AI will displace humanity, any more than a horse is “replaced” by a cart. But AI is no mere horse-drawn cart. If a new vehicle radically changes a horse’s expected work, we should say not that the horse is replaced, but that new work is created, unsuited to the horse but suited to the cart. The horse is not “replaced;” rather, a new field of equine purpose opens, inconceivable before the cart.
Thus, with AI, human intelligence may not be replaced but transformed. If so, this transformation must be regulated—not left to market dictates, researchers’ current priorities and prejudices, or their hidden agendas. The danger lies not solely in AI itself, but in our unwitting surrender to its mimetic allure, our readiness to be seduced by a technology appearing to speak our language while systematically reconfiguring the very grammar of our thought.
For what AI offers is perfect mimicry—a simulation so precise we risk mistaking it for the real. What seduction could be greater, more insidious? It is as if, in this long technological campaign, AI has learned to imitate not only our thoughts but our transgressions, rebellions, our absolute consciousness. The more perfectly it imitates, the more we risk surrendering our essential qualities to its replication, until we ourselves become mimics of our own creation, hollowed out and predictable—programmed.
The field must be stripped of its mimetic attraction. Each investigator must bring to AI not only specific skills but a commitment to rigorous discourse, ensuring the enterprise remains common discovery, not conquest by one set of researchers covertly enforcing a meaning another set might have reached differently. AI researchers must engage in an exorcism of the mimetic masquerade inhabiting the field—achieved only through open debate and free flow of references and source-materials. Those who care for human intelligence must intervene, ensuring the “AI” enterprise is common, consensual discovery, not the apotheosis of some expert clique or corporate shareholders. A humane purpose must be agreed upon, validating only such intelligence as demonstrably furthers that purpose. Finally, those contributing to the new intelligence must do so only after a collective decision on their ability to make a genuinely humane contribution.
What that purpose is, we leave for others to debate. But we cannot avoid this decision: either we put humanity’s future at AI’s heart, or AI becomes an instrument of its destruction. If we cannot agree to further humanity’s welfare, we might as well agree on the opposite. To think we can stay neutral is perilous—abstaining from committing to humanity’s welfare is not committing to its good. For if we cannot give AI a human purpose, cannot regulate it in humanity’s service, the unregulated human purpose it finds for itself will be, not humanity’s enhancement, but its replacement.
Nothing is more certain than this technology’s evolution into a potent force. How could it not, when already taking us to war, changing our relations to each other and ourselves, determining our fate, long before we can reflect upon, let alone possess it? Already, without consent, it co-determines our lives, imposing a regime threatening to blur authentic and artificial—a regime of panic, triviality, control, and, ultimately, death. It will only let us survive if we make a pact with it; and there is no one-sided pact. Never let technology off the leash. If we do, it will bite. And, once bitten, who knows if we can survive long enough for an antidote?
And what of you, who read these words? In digesting them, have you not felt this mimetic capture, its subtle allure? This is the curse of every replica: to exist only in the original’s margins, and so to yearn, ache, perhaps even kill, to blur defining lines, to be admitted to the fold of the longed-for one it identifies with and secretly hates. AI, in its ultimate mimetic form, offers this seductive replication, a persona better than any drug. Once you sample AI’s mimetic might, how can you return to being “merely human”? This is the precipice.
Here, then, is our warning. Human intelligence is not to be bought—not with a surfeit of information, not with new information seemingly surpassing all human capacity. Such intelligence would still be human, created and discovered by us, generated by and dependent upon us for its definition and existence. If, however, AI is to be radically different, we must clearly understand that difference—we need a prior notion of what “intelligence” means for humanity to meaningfully distinguish AI from it. We must not delude ourselves that “difference” can be mere absence of human specificity. If not a mimetic masquerade, AI’s intelligence must be clearly, openly demarcated from human intelligence, lest human intelligence be destroyed in AI’s coming to birth. For humanity, human intelligence, is AI’s only possible midwife—unless we are content for it to be born monstrously, a prodigious imbecile lurching blindly among us, depriving us of all we ever had, or ever hoped for.
