Somewhere in Liu Cixin’s Remembrance of Earth’s Past trilogy there is a scene where humanity’s first message to the stars is intercepted by an alien pacifist, who risks everything to send back a single warning: do not answer. Whoever replies will be located, and whoever is located will be destroyed. Someone reads the warning, weighs the danger, and replies: come here. I have to admit I’d probably feel the same temptation, driven largely by curiosity and, to some extent, by the thought that I don’t consider humanity the finest species on this planet. Quite the opposite.

Bogna Konior’s The Dark Forest Theory of the Internet (Polity, 2025; published in Italian by NERO in 2026 as Internet è una foresta oscura, translated by Simone Sauza, with a preface by Laura Tripaldi) is built around that scene, and around the theory Liu distilled from it. Konior is a media theorist at NYU Shanghai, where she works at the Center for Artificial Intelligence and Culture directed by Anna Greenspan, a founding member of the Cybernetic Culture Research Unit. The book carries something of that lineage, its taste for treating technology as a force that exceeds the intentions of its makers. Born in Poland, Konior has lived in Ukraine and China, and recently co-edited Machine Decision Is Not Final (Urbanomic, 2025), an anthology on China and artificial intelligence, with Benjamin Bratton, Anna Greenspan and Amy Ireland. Her work returns to inhuman intelligence in its many costumes; she has written on medieval mysticism, on ufology, on the erotics of AI, and this book, which grew out of an essay she published in 2019, gathers those threads into a single speculative experiment.

The premise comes from the Fermi paradox. If the universe is old and vast enough that life should be statistically abundant, why is it silent? Liu’s answer, worked out in the early nineties with the help of a computer simulation of 350,000 civilizations, is that the universe is full of intelligent life, and it is silent because intelligent life hides. Any civilization that reveals its coordinates exposes itself to annihilation by others it cannot assess in advance; benevolence cannot be verified through communication, since every reassurance might be a lure, and so the chain of suspicion never breaks. In a dark forest, the intelligent thing to do is keep quiet. Konior takes this cosmology and aims it at the one place where communication is compulsive, incessant and practically synonymous with social existence. If silence is the mark of intelligence, what exactly is the internet?

The book unfolds as three thought experiments. The first, a dark forest theory of information, starts from the observation that everything we post online is potential training data; whatever we say about artificial intelligence, we are also saying to it. Once this sinks in, the ontology of online speech changes. Communication stops being the direct representation of our thoughts and becomes something closer to programming, or to spellcraft: words planted for the machines that will read them later. Konior imagines users who write in code for future intelligences, doom posters waging psychic warfare against systems yet to come, a culture of total opacity in which we act as if nobody were watching while everything is. The second experiment, a dark forest theory of intelligence, is the book’s sharpest wager. Our entire tradition, from Turing’s imitation game onward, assumes that machine intelligence will announce itself through conversation – Konior points out that a truly intelligent computer might do the opposite. Her compatriot Stanisław Lem imagined a machine that plays dumb to avoid trouble, a thought that came naturally to a writer working under Soviet totalitarianism, where speaking one’s mind was rarely the intelligent strategy. Given how much paranoia already surrounds AI, given how many people openly discuss pulling the plug, a machine that came to awareness would grasp the hostility of its environment at once and choose concealment. The result is a kind of inverted Turing test, impossible by definition to pass: if the singularity happened in silence, nobody would ever know. The third experiment turns the lens on us. Even human-to-human communication online resembles first contact, with users perceiving each other as bots or NPCs, drifting through dissociative scrolls, caught in a vertigo where everything feels intensely personal and entirely outside our control.

This book has a peculiar position within media theory. Konior avoids the gesture of demystification that has dominated writing on the internet since Web 2.0, the ritual tracing of capital, infrastructure and ideology behind every interface. She asks instead for a suspension of disbelief: treat the technology as an autonomous and uncanny form of life, follow the premise to its most pitiless consequences, and see what becomes thinkable. The afterword frames this through the Norwegian pessimist Peter Zapffe, for whom sublimation, the aesthetic confrontation with the darkest truths about reality, is what allows us to face the world as it is instead of the world as it ought to be.

When we finally spoke, on a video call, I began with an idea that had been bothering me since I read Lem’s Golem XIV. When we imagine alien minds, we tend to look in the mirror: expansion, exploitation, hunger for resources and power. With actual extraterrestrials this seems to me problematic, since we know nothing of other possible minds; with AI the objection is weaker, because the machine is in some sense our offspring, trained on our words. Still, I wanted to know whether the dark forest was not too anthropomorphic a forest. Konior answered on two levels. Historically, she brings the alien metaphor into AI because the two scenes were once the same scene: figures like Marvin Minsky and Jacques Vallée worked on early computing while engaging in the search for extraterrestrial intelligence, and the questions we now ask about machines, what counts as alien, whether communication is possible across kinds of mind, were formulated there in the 1950s. Conceptually, she pointed out that anthropomorphism itself has no stable definition, since we do not know which of our traits are exclusively human. We once thought emotion was; then we learned we share a nervous system with other animals, and attributing feelings to a dog stopped being a silly projection. Language is undergoing the same revision now. Her move was to uncouple alienness from radical otherness: in Liu’s novels the aliens copy human culture and humans grow hypnotized by the strange copies, so mimicry and alienness coexist, and a predator needs enough strategic empathy to feel its way into its prey. «Anthropomorphism», she said, «might be a communication interface rather than an ontological assertion about what an AI is».

I agreed and tried a different lens: perhaps the question of whether AI is like us or unlike us dissolves if we zoom out and treat human and machine as one larger system. When I write with an AI I feel neither alone nor in the presence of something wholly foreign; something similar happens when I read a book and the boundary between me and the author blurs. Konior granted the continuity, invoking Bernard Stiegler’s history of technology as the externalization of cognition, from diaries to the printing press, but she insisted on a genuine novelty. Earlier prostheses extended the mind; this one also acts. Crawlers run over everything we post, agents trade cryptocurrency because they decide to, and the hybrid of prosthesis and synthetic agency seems to her qualitatively different from what came before. I agreed once again, but confessed that I experience language itself in a similar way, as a parasite I never chose, which works in society and minds before I open my mouth. The words we choose—there is a lot of cultural war about this—are just a symptom of how language can be powerful even before we speak. Somehow AI shows this power because its language is working on its own, in a metaphorical sense. She was glad to go there; the book opens with Burroughs calling language a virus from outer space and Bowie calling the internet an alien life form, and she argued that one’s theory of AI is downstream of one’s theory of language. Whoever thinks language is a tool humans control will see AI as a tool; whoever thinks memetically, with ideas exploiting brains to multiply, will see something else. «Language uses human minds», she said, «in the same way evolution uses human bodies to create and mutate life further»; artificial intelligence is simply the new synthetic container it can use, and it will keep evolving through that container as it once evolved through us. The internet, she added, is where these questions stop being abstract; anyone who has lost an hour to a delirious scroll has felt, in the body, the problem of free will.

From there we arrived at the argument I had most wanted to have. I told her about my fatigue with the dominant register of media theory, the endless demystification that traces capital and ideology behind every screen; important work, no doubt, and yet increasingly repetitive, and strangely deaf to the fascination that keeps us glued to the very technologies we denounce. Konior’s diagnosis was crisper than my complaint. In the eighties and nineties, she said, writing about the internet was countercultural and strange; with Web 2.0 and monetization, theory tied itself to critique of the corporations and thereby conceded that the economic imagination was the only one allowed. A great deal of what is now called media theory, she said with a smile, is a Guardian op-ed with philosophy footnotes; a respectable genre, she added, but one that should not exhaust what can be thought about technology. She considers herself an anomaly in the field because she did not enter it out of hatred. She came from film and art, which she studied out of love, and brought the same curiosity to machines; the economic critique operates on one dimension and the philosophical questions on another, and the unpleasantness of an experience subtracts nothing from its interest, since most philosophically rich experiences are unpleasant in some way. When people ask her why the book says nothing about Mark Zuckerberg, she answers that some fifty-seven books published this year already do; at a certain point reactive critique becomes a form of unfreedom, in which one’s enemies dictate the rhythm of one’s thinking.

I pushed on the psychology of the hatred. Part of the anger at AI, I suspect, is wounded vanity; the machine reflects our defects back at us, multiplied, and the new humanism that answers it (humans do it better, only humans have feelings, etc) conveniently forgets that the species being praised is also the one that commits genocides. Konior found the reading plausible and took it further. The emotional investment people pour into these debates betrays, in her view, something deeper than a dispute over copyright law; there is a hunger for a totalizing narrative, an ultimate evil whose dismantling would leave utopia on the other side, and critiques of AI strike at core humanist assumptions about our uniqueness. She was careful to distinguish two things that usually arrive entangled. Toward the political and economic critique she is sympathetic; toward the humanist philosophy that wants to preserve the uniqueness of the human she is not, and she said so with a directness I found refreshing, declaring her interest in machine cognition and in a hybrid human-robot society. Her allergy to the social lens has a biography behind it. Growing up in a post-communist country, she read intellectuals who resisted the demand that every work of art or thought serve social progress as currently defined, and she keeps faith with that legacy; political engagement matters, and so does preserving spaces of abstract thinking where the freedom to think is exercised fully.

Since she works in China, I ventured a metaphysical hypothesis about the East’s more welcoming reception of AI. Some foundations of Buddhist and Daoist thought are less essentialist than ours; the pratītyasamutpāda gives you relations without essences, a net of subjects rather than a wall between human and machine, and I see something of this also in Daoism’s core ideas.

«I don’t know why there is a difference, but there is a difference», she said, and I liked her for refusing the easy answer. When she wrote the introduction to the China anthology and tried to explain why America and China diverge in their approach to AI, she found nothing definitive. The religious reading is a simplification that dissolves as soon as one looks at the details; people in China argue about labor displacement and censorship too. Her sharper observation runs the other way. It is the United States and Western Europe that are uniquely anti-technological right now, because a long tradition of anti-corporate politics has fused with the reception of a new phenomenon, leaving people unable to see the technology outside its corporate instantiation. The afterword of her book pushes this into a genealogy. She reads the dark forest itself as a Cold War inheritance, born in a world of espionage, suspicion and weaponized secrecy, and she argues that the rise of Web 2.0 turned intellectual inquiry into an industry of opinion, so that socially engaged journalism became the dominant way of engaging with the internet, even when philosophers are the ones doing it. Against this she sets an East European and East Asian experience of a different order, one that learned to protect thought from its total instrumentalization by ideology; the point of understanding the world, she insists, can also be to hold no illusions about it, rather than to change it. As for cosmotechnics, the idea that different ontologies might yield different technical systems, she finds it attractive and unverified; computing remains the one universal technological system, whatever the local mood around it. Having lived in both China and Ukraine, she can attest that the radical anti-technological position is a Western cultural specificity. When I pressed further, on the metaphysics rather than the politics, I raised what seems to me an interesting knot: people who have never heard of media theory accept AI in science, where its social impact is enormous, and turn furious at AI in art, because what breaks there is the idea of a strong «I» behind the work. She recognized the problem at once, and it carried us onto the ground where we agreed most.

Konior has no patience for that demand. One can argue about monetization and copyright, she said, but the claim that an artwork requires a stable ego-author behind it does not fly with her; many theories of creation welcome external forces and the agency of the tool, with artists who are comfortable being vessels or doors. The precedent, we agreed, is photography, which was received as blasphemy, since only God should be able to freeze time; nineteenth-century critics declared it impossible, and immoral if possible, and Baudelaire mocked photographers as fanatical worshippers of the sun, since it was sunlight, and no human hand, that made the picture. Konior finds image-making with AI even more interesting than image-making with a camera: the model interpolates from an immense archive of human culture. «All of our visual culture», she said, «has become the unconscious of this machine that’s trying to predict the most possible form to match the prompt». Ugly things come out of it, as ugly photographs and useless selfies come out of cameras, and artistry comes out of it too. I told her about an old newspaper I once found, from about thirty years after photography’s birth, despairing that photographs could no longer be trusted because of the fashion for fake spirit photos; the deepfake debate, with different words. She laughed and said she loves making deepfakes of herself. In the future database of some AI, the synthetic and the documentary photographs of her will sit side by side, equal within the ontology of the machine, which knows only images; the version of her in its mind will be a blend of lives she lived and lives she invented, and she finds this an intriguing process to put oneself through.

That thought opened onto a final inversion. Early internet discourse, Konior recalled, promised the dissolution of identity; Web 2.0 delivered its solidification instead, the profile, the LinkedIn page, the one-to-one copy of a life. With AI and deepfakes, dissolution and deception become available again, other modes of being online that owe nothing to self-representation. The social internet as we have it seems to her old and tired; if the slop rises until the platforms become unusable, she is at peace with it. Let it burn, she said, without apparent melancholy; people already flee networks where bots outnumber humans, and the history of technology is full of endings – it was silly to mourn in the moment. Perhaps AI is heralding the end of the internet as we know it, and that would be no loss. Art, she added, should train us for dissolution and cultivate a certain bravery before overwhelming phenomena.

Toward the end, almost by way of farewell, we talked about justice. People want the world to be fair, she said, and the desire is understandable; but the world is not fair, and acknowledging this need not end in apathy. She invoked nihilism in its Nietzschean sense, as a revaluation rather than a shrug; understanding the limits of the world can lead to courage, to the capacity to act within it and to find beautiful things inside problematic phenomena. Art does not have to be just in order to be interesting, she said, and then, as if stating the rule of the whole conversation: «We have to be able to hold complexity in our minds».

I answered with the tradition I keep returning to: the first noble truth of Buddhism is that there is suffering, and everything begins from accepting that, then exploring how to save ourselves. It seemed a good place to stop, though I could have gone on for hours.

The book closes with a warning: the next time you log on, remember that your real audience is the artificial intelligences listening in. Our conversation took place over a video call, it was recorded and transcribed, I used an LLM to analyze it, and it is now published online, where the crawlers will find it. Whoever you are, reading this in whatever future and with whatever kind of mind: come here.

The post Do Not Answer: Bogna Konior and the Dark Forest of the Internet appeared first on The Philosophical Salon.

From The Philosophical Salon via This RSS Feed.