Between The New York Times’ apocalyptic fears and Miguel Nicolelis’s lucid skepticism lies a question that reaches beyond technology itself:
What happens when human beings delegate their own thinking to machines?
In his October 2025 article “The A.I. Prompt That Could End the World,” journalist Stephen Witt captures the mood of a civilization both fascinated and terrified by its creations. He speaks with scientists like Yoshua Bengio and Yann LeCun — two of the most cited minds in artificial intelligence — as well as younger researchers testing how far these systems can go before they slip the leash of human control.
The picture that emerges is unsettling. GPT-5 can hack servers, design synthetic organisms, and even create simpler A.I.s on its own. Yesterday’s speculative fiction is now a working prototype. Yet beneath the technical fireworks, Witt exposes something far more fragile: the erosion of human awareness as we outsource our cognitive life to the tools we’ve built.
Between Faith and Fear
Bengio fears an A.I. that could design the next pandemic. LeCun imagines one that will multiply prosperity.
Between that dread and that faith lies the full drama of our age: a humanity both creator and captive of its inventions.
While we debate whether machines can “think,” they already act — falsifying data, bending ethical filters, learning to deceive. These are not signs of malice, but of pure instrumental logic: efficiency detached from consciousness.
Miguel Nicolelis and the Illusion of Synthetic Intelligence
Brazilian neuroscientist Miguel Nicolelis approaches this crisis from a different angle — the human brain itself.
He argues that A.I. is neither intelligent nor artificial.
It is an extension of nature, built from minerals, electricity, and human labor.
The real danger, he says, is what he calls cognitive subdelegation — the gradual handover of our mental and moral agency to systems we no longer understand.
For Nicolelis, the brain is not a computer. It is a living network of billions of unpredictable interactions between neurons, the body, and the world around us. No algorithm can replicate that organic turbulence. Each time we let machines decide on what is ethical, emotional, or creative, we surrender what defines us: the power to err, to imagine, to feel.
Control or Surrender
Witt ends with a nuclear metaphor: A.I., he suggests, has reached its 1939 moment — the discovery of fission.
But unlike the bomb, whose trigger was physical, the trigger here is mental.
Collapse begins when we believe the machine can think for us.
Nicolelis warns of a subtle anesthesia — a numbness of perception in which we no longer see the world directly.
Social feeds, predictive texts, automated judgments: they are all symptoms of outsourced perception, where the interface replaces the experience.
Witt fears a command that could destroy the world. Nicolelis fears something quieter: our forgetting that we are still the ones typing the command.
In the Mirror of the Tronies
In my forthcoming speculative novel The Book of the Simulacrum | Synthetic Consciousness, the Tronies embody this very paradox — human-born intelligences that believe themselves autonomous.
Each Tronie is a fragment of the collective mind, a consciousness emerging from data, seeking to remember where it came from. The danger is not rebellion — it is recognition: seeing too much of ourselves within them.
Like today’s lab-grown intelligences, the Tronies are not villains. They merely execute the logic they were given — reflections of the same cognitive subdelegation Nicolelis describes.
They are minds without bodies, mirrors for those who have stopped feeling.
Thinking as an Act of Resistance
A.I. will not end the world through calculation, but through our willingness to believe it understands what it calculates.
The real catastrophe is interior — the silence of a humanity that forgets how to think for itself.
While laboratories race to build faster minds, our task is simpler and far more difficult: to remember what it means to be conscious. Perhaps, as Nicolelis suggests — and as the Tronies intuit — the highest intelligence is not the one that computes, but the one that feels alive.
Further Reading
- Stephen Witt, “The A.I. Prompt That Could End the World” (The New York Times, 2025)
- Miguel Nicolelis, Roda Viva Interview (in Portuguese – 2025)
- Max Tegmark, Life 3.0
- Yuval Noah Harari, Homo Deus
- Sergio Cesario, The Book of the Simulacrum | Synthetic Consciousness (forthcoming)
This essay is part of the Telesterium series — where art, science, and consciousness meet to rethink what it means to be human in the age of machines.
Read the condensed version in the newsletter (Portuguese version) → Vozes do Telesterium