The key problem with the “brain in a vat” thought experiment
Even with the best technology imaginable, you'd probably never be able to exist as a consciously aware brain in a vat.
by Adam Frank
This article was first published on Big Think in 2024. It was updated in March 2026.
There is a certain kind of story about nature and our place in it that says all we experience, all we feel, is nothing but neural circuitry. According to one thought experiment, if I were a smart enough doctor or computer scientist, I could put your brain in a vat, hook it up to a bunch of electrodes that enable some kind of simulation of the world, and you’d never know the difference. Fed the proper electrical stimulus, your brain would conjure up the exact same experiences as if you were wandering around in the world in your body.
But is this “brain a vat” thought experiment right? Is direct experience — that most intimate and immediate sense of life and the world — really reducible to the wiring of neurons? The answer, I believe, is not at all. More importantly, the mistake at the heart of this idea is a primary example of the inability of certain philosophical perspectives on science to confront the centrality of human experience in doing science.
Reducing conscious experience
The thought experiment is prevalent in modern culture, inspiring numerous science-fiction films and TV shows, like The Matrix (so awesome) to Altered Carbon(very cool, but less awesome). It’s also the underpinning of much-talked-about ideas like the simulation hypothesis or Boltzmann Brains. Taken as a whole, brains in vats (let’s abbreviate them as BIVs) exemplify a particular metaphysics of science called reductionism, in which everything we experience can be reduced to the activity of lower levels in the hierarchies of structure in the physical world. Reductionism plays a big role in The Blind Spot: Why Science Cannot Ignore Human Experience, a book that Marcelo Gleiser, Evan Thompson, and I published in 2024. In it, reductionism is seen as one of a constellation of views through which direct experience gets forgotten as that which makes scientific activity possible. The most dramatic feature of Blind Spot’s philosophy is when useful abstractions — extracted from scientific practice — are substituted as being more real than the rich, seamless life-world that is experienced. That’s exactly what happens with BIVs.
There are a number of ways to see how the BIV idea gets things wrong, but today I am going to draw from a lovely paper that my Blind Spot co-author Evan Thompson wrote with Diego Cosmelli. It’s called “Brain in a Vat or Body in a World? Brainbound versus Enactive Views of Experience.” Thompson and Cosmelli want us to take the BIV thought experiment seriously. What would it take to actually “envat” a brain? Any answer to this question — even one that assumes we had the best technology imaginable — must wrestle with a thorny problem: the fact that the brain and body are extremely interconnected, perhaps inextricably so.
Designing a BIV
Given that brains are just one organ in the collection of systems we call an organism, we would need to provide our brain with an apparatus that keeps it alive and functioning. Your brain would need to be constantly fed various substances at their exact concentrations — at the same frequency and in the same brain locations — to maintain healthy function. So, we would need a highly specific, highly resolved circulator system to deliver the required substances. Note, however, that these substances are not passively delivered: It is brain activity itself that determines where, when, and how the circulation system should carry out its task. This means our artificial circulatory system must be coupled at the highest resolution to the brain and its own activity in self-regulatory loops. As Thompson and Cosmelli put it: “Our life-sustaining system must support this intrinsic activity and respond to it locally and systemically at any given instant, independent of any outside evaluation of the brain’s needs.”
There is more, however. As a living organ, the brain is not just maintained from the outside, or exogenously. The brain controls motor functions: how our body parts actually move, including our sense organs. So, instead of one-way control, there’s a continuous sensorimotor loop: endogenously controlled movement changes how the senses are stimulated, which gives rise to new movement. The brain has to deal not just with what is happening in the outside world but with what is happening as a result of its own intrinsic activity. This activity is highly nonlinear: It’s not a straightforward result of whatever the sensory inputs are, and it all has to be kept within viable limits so it doesn’t disrupt the homeostasis of the healthy, functioning brain.
Taking these and other requirements together, Thompson and Cosmelli conclude that to really envat a brain, you must embody it. Your vat would necessarily end up being a substitute body. Note that they aren’t claiming the substitute body has to be flesh and blood. Instead, they demonstrate how the BIV thought experiment undermines itself. Its fundamental idea is that neural circuitry is somehow the minimal condition/structure needed for experience. Instead, Thompson and Cosmelli demonstrate that being in a body that is itself active in the world is the minimal condition/structure necessary for experience. By beginning with the science of brains as living organs in living organisms, they demonstrate one way that the BIV idea undercuts its own logic. In other words, brains may be necessary for experience, but they aren’t sufficient. They are one part of the holism that is embodiment in a world — the true “seat” of experience.
“We’ve given reasons to think that the body and brain are so dynamically entangled in the causation and realization of consciousness as to be explanatorily inseparable,” write Thompson and Cosmelli.
To close, I want to note that these kinds of arguments are not just an academic philosophical game. There’s a steady drumbeat these days of people making powerful claims for AI. Some of these claims draw directly from philosophies animating the BIV argument. Understanding exactly where the flaws in that argument appear is one step to making sure we don’t end up building a deeply flawed society that rests exactly on those flaws.
Yes but what if such a brain apparatus was ALSO connected to an advanced sensor system letting it interact & communicate with “the world!?” Maybe the internet as well! Bc clearly just a consciousness confined to a vat would likely go mad! The immorality cruelty aspect must also be as prime consideration! Go to Amazon.com and ck out “Jupiter’s Dominion!” by Mj deyoung! It’s a short inexpensive sci-fi story download that addresses this issue as part of the plot!! Please post a review!
Adam Frank’s attack on “brains in vats” is, frankly, a strawman. The original thought experiment is not an engineering proposal about how to keep a brain metabolically viable; it is a phenomenological and epistemic device that simply stipulates a perfect proxy for all the brain’s usual inputs and outputs. Once you grant that stipulation, questions about metabolism, circulation, or “how hard it would be to build the vat” are, by definition, irrelevant to the point.
By shifting the discussion to biological implementation, Frank quietly changes the rules of the thought experiment instead of engaging its payload: that if the total informational state of the brain is identical, then the experienced world is indistinguishable from the inside, regardless of whether the data stream originates in a body or a computer. That is why his argument doesn’t refute the Brain in a Vat scenario; it replaces it with a different question altogether.
The most significant failure of the article is that Frank's "Blind Spot" is actually his own. He is so committed to a "Brain-Bound" biological view that he cannot conceive of consciousness as an entity that interacts with a data stream.
Frank is arguing against the wrong thing. Brain in a Vat is not mainly about the biology of the brain or how you would engineer the setup. It is about first-person experience. If the experience is identical, then it is identical from the inside, no matter where the input comes from. And because of that, the thought experiment is not automatically reductionist. It is also compatible with non-reductionist views of consciousness.