Epicurus had important things to say about happiness, pleasure, fear, and anxiety. But before the modern world reduced him to the status of a street-corner therapist, his most important contributions to human advancement were understood to be in a very different field — that of understanding reality, and the true nature of things. The ancient world recognized that Epicurus's account of how knowledge is possible was among the most significant and original contributions any philosopher had ever made. It is time to recognize that again.
The reason is not merely academic. Advanced artificial intelligence technology is making it increasingly difficult to trust what we see and hear in the world around us. Audio and video can be generated that is indistinguishable, to direct observation, from recorded reality. Images of people, events, and places can be fabricated that human senses cannot reliably distinguish from genuine photographs. There is little doubt that as the technology matures, the generation of convincing synthetic taste, touch, and smell experiences will follow. We are entering a world in which the question "can I trust what my senses are telling me?" is no longer a philosophical puzzle reserved for academic discussion. It is a practical daily challenge.
When we think about what our response to this new world will be, we can predict three broad categories of reaction. The first is despair: no sensations can be trusted, and there is no way in any given situation to be certain whether what the senses are reporting is true or false. The second is the claim to a privileged criterion: some sensations are so clear and sharp, or some authority so reliable, that they are unmistakably true. The third is confidence based on understanding: all sensations can be trusted to report what they report, and what our minds do with those reports is where the question of truth and falsehood actually resides.
These three positions were identified and argued in the ancient world more than two thousand years before AI existed. Their broad outlines remain the only available options today. And the one that was right in the ancient world — the one that Epicurus developed against both the Skeptics and the Stoics — is the one that offers genuine guidance in ours.
The first position — that no sensations can be trusted — was the position of the Radical and Academic Skeptics: the heirs of what Socrates and Plato had introduced, extended through Pyrrho of Elis and eventually systematized by Sextus Empiricus in the second century AD. The Academic Skeptics, working from within Plato's own Academy, argued that no sensory impression could carry its own guarantee of accuracy — and that therefore the only honest response to any question was suspension of judgment. The Pyrrhonists pushed this further, extending suspension of judgment to the skeptical claim itself, producing a position of total philosophical paralysis.
In an AI world, this position maps directly onto the response of those who conclude that because synthetic media cannot always be distinguished from genuine media, nothing can be reliably known from sensory experience. Every report, every video, every image, every witness account must be suspended pending some standard of verification that is never itself secure. The result is the exact paralysis that Epicurus identified in the ancient Skeptics: a life without the possibility of confident judgment, without urgency, without the capacity to act on reliable knowledge. A world in which AI has convinced us that nothing can be known is a world in which the Skeptics have won — and Epicurus understood that the Skeptic victory is not philosophical honesty but philosophical suicide.
The second position — that some sensations are so clear they can be trusted — was the position of the Stoics, who developed their theory of knowledge around what they called the "kataleptic impression." This was an impression so clear and distinct, the Stoics argued, that it carried its own internal guarantee of accuracy. The Skeptics' most devastating argument against the Stoics — that no impression can satisfy this standard because hallucinations, dreams, and deceptions produce impressions qualitatively identical to genuine ones — applies with equal force in the AI context: if artificially generated images are perceptually indistinguishable from genuine ones, the criterion of clarity cannot separate true from false. The kataleptic impression is not a standard that survives the challenge of AI-generated media.
The same logic applies to any modern equivalent of the Stoic position: the claim that some authority — a government certification, a platform's content moderation, a blockchain verification, a chain of custody document — can sort genuine from synthetic on our behalf. Whatever the authority, it is itself capable of being deceived, fabricated, or corrupted. Delegating the problem of knowledge to an external certifying authority is the modern equivalent of delegating it to the Stoics' dialectical logic — it multiplies the problem rather than solving it. This was always the Epicurean critique of the Stoic approach: its obsessive resort to dialectic, syllogisms, and formal logic produce results that are as unsatisfactory as they are inaccessible to ordinary people.
The third position — that all sensations can be trusted to report honestly what they report, and that truth and falsehood reside in the mind's interpretation of those reports rather than in the sensations themselves — is the Epicurean position. It is the one that remains standing after both the Skeptic and the Stoic have been refuted. And it is the one that offers genuine guidance in the age of AI.
The foundation of Epicurus's account of knowledge is stated directly by Diogenes Laertius, drawing on Epicurus's lost work The Canon:
In The Canon Epicurus states that the sensations, the preconceptions (prolēpseis), and the feelings (pathē) are the criteria of truth. (Diogenes Laertius, X.31)
These three natural faculties — the five senses, the anticipations built by accumulated experience, and the feelings of pleasure and pain — are what nature has provided to every living creature from birth as the instruments for navigating reality. They are not the result of philosophical training or rational discipline. They are given. Philosophy's task is not to replace them with something more sophisticated but to prevent false opinion from corrupting them.
The central claim about sensation is both radical and precisely formulated. Sensation, Epicurus held, is always true — but this does not mean what its critics have taken it to mean. It does not mean that every impression accurately represents the external object that caused it. It means that sensation, as a physical process, reports honestly and accurately what it receives. As Sextus Empiricus records the Epicurean position:
Sensation, being perceptive of the objects presented to it and neither subtracting nor adding nor transposing — being devoid of reason — constantly reports truly and grasps the existent object as it really is by nature. (Sextus Empiricus, Against the Logicians M VII.369, drawing on Epicurean sources)
The key phrase is "neither subtracting nor adding nor transposing." Sensation does not interpret, select, or rearrange what it receives. It is, in Norman DeWitt's formulation in Epicurus and His Philosophy, "irrational" — not in a pejorative sense but in the precise sense that it operates entirely below the level of rational judgment. And precisely because it does not judge, it cannot misjudge. Only the mind, when it adds its own opinion to the raw report of sensation, introduces the possibility of error.
Diogenes Laertius records the same principle from the Letter to Herodotus:
All sensation is irrational and does not admit of memory; for it is not set in motion by itself, nor, when it is set in motion by something else, can it add to it or take from it. (Diogenes Laertius, X.31)
This is the technically precise Epicurean answer to the bent oar and the tower that appears round from a distance. The oar in the water looks bent — and that sensation is true. It accurately reports what occurred at the visual interface of eye, oar, and water given the laws of light refraction. The error is not in the sensation but in the hasty judgment that therefore the oar is bent in fact. Correction comes not from abandoning sensation but from more observation and better judgment: pick up the oar, look again, test the report against further experience. As Principal Doctrine 23 states: "If you fight against all your sensations, you will have no standard to which to refer, and thus no means of judging even those sensations which you claim are false."
Of the three natural criteria — sensation, anticipations, and feelings — the least discussed in modern treatments of Epicurus is the second: the faculty he called prolepsis, translated into Latin as anticipationes and into English most naturally as "anticipations."
Anticipations are the generalized pre-concepts and pattern-recognitions that the mind builds from repeated sensory experience. When you have encountered horses many times, your mind has assembled those experiences into a recognizable pattern — a "preconception" of what a horse is — that allows you to recognize a new horse immediately without having to process each feature from scratch. This is not a rational construction or a definition arrived at through dialectical method. It is an automatic, empirically grounded recognition built by nature through the accumulation of experience.
Epicurus insisted that this faculty is built into every normal human mind from birth. It does not require philosophical training or the study of geometry. The ploughman who has observed horses his whole life has anticipations of horses just as reliable as the philosopher's. Diogenes Laertius records explicitly that Epicurean prolēpseis are formed entirely from repeated sensation — they are empirically grounded at every level (X.33). This is the precise point at which Epicurus separates himself from Plato: Plato grounded certain concepts in the soul's pre-birth encounter with eternal Forms, making a priori knowledge possible. Epicurus grounded all concepts in accumulated sensory experience, making knowledge something available to every human being rather than reserved for the philosophically initiated.
The faculty of anticipations is exactly what is at work when we navigate the problem of AI-generated media in practice. We are not helpless in the face of convincing synthetic content because we do not rely only on the sensation of the moment. We rely on accumulated experience — the patterns we have assembled from thousands of real encounters, conversations, and observations — to evaluate whether what we are being presented with coheres with the larger fabric of our experience. The synthetic video that shows a political figure saying something they have never said is detectable — sometimes immediately, sometimes only on reflection — not because the pixels carry their own certification of falsity but because the claim embedded in the video conflicts with the accumulated anticipations we have built from real experience with that person, with the surrounding context, and with the natural world.
The most practically important insight in the Epicurean Canon for navigating an AI world is this: error does not live in the sensations. It lives in the opinions the mind adds to the sensations.
This sounds simple but its implications are radical. The Academic and Pyrrhonist Skeptics concluded from the possibility of deception that sensations cannot be trusted. Epicurus's response was that they had identified the wrong culprit. The sensation of the AI-generated video is real — your eyes and ears are accurately reporting the pixels and sound waves that reached them. The sensation itself did not deceive you. What deceives you — if deception occurs — is the hasty opinion that the sensation constitutes adequate evidence for the conclusion "this video is real and accurately represents an actual event."
The Epicurean Canon tells us that the correction for this error is not to distrust sensation but to be more careful about what the mind adds to sensation's report. Does the content of this video cohere with what your accumulated anticipations tell you about this person, this situation, this context? Does it survive scrutiny when tested against multiple independent observations? Do the feelings it produces — the unease, the implausibility, the sense that something doesn't fit — deserve attention as genuine signals rather than being suppressed by the excitement of a striking claim?
Principal Doctrine 24 gives the basic methodological principle: "If you reject absolutely any sensation and do not distinguish between the opinion about what awaits confirmation and what is already given by sensation, by the feelings, and by the mind's observations, you will throw the remaining sensations into confusion with your foolish opinion, and thus destroy every standard for evaluating things."
The point could not be more directly applicable. In an AI world, abandoning sensation entirely — concluding that nothing can be known from experience — is not epistemological sophistication. It is exactly the error Epicurus identified in the ancient Skeptics. It destroys the standard by which anything could be evaluated at all.
As we move deeper into a world where artificial intelligence can produce synthetic sensory experiences of rapidly increasing fidelity, the Epicurean perspective becomes more valuable rather than less. And the reason is precisely that its foundation is not in the fidelity of any particular medium — not in 35mm film, not in digital photography, not in certified blockchain records — but in the natural faculties that every living human being has possessed since before any of those technologies existed.
The standard of truth for Epicurus is not, has never been, and will never be resort to supernatural authority, ideal forms, Aristotelian essences, or any artificial philosophical construction. It is the continuous and repeated testing of actual experience through the sensations, the anticipations built from accumulated experience, and the feelings of pleasure and pain that nature has provided as the honest signal of how things are going.
Most of us are already unable to judge reliably by direct inspection whether a given video was recorded or generated. But we can also begin to recognize something that Epicurus would have found obvious: the medium through which an experience is delivered is not the determinative question. What matters is whether the content of that experience coheres with the accumulated pattern of real encounters with the world. AI-generated media is, in the relevant sense, just as "real" as media recorded on any earlier technology — in that it presents genuine sensory inputs to our genuine sense organs. What it is not is necessarily a reliable report of actual events. The evaluation of that question belongs not to the sensation itself but to the anticipations and judgment of the mind.
This has always been the situation. The ancient forgery of a document, the false witness in a trial, the sophist's manipulation of language — these were the AI-generated media of their age. Epicurus's point was never that deception is impossible. His point was that the tools for detecting deception are natural, universal, and available to everyone — not reserved for the philosophical sage, not dependent on supernatural revelation, not requiring the mastery of dialectical logic. They are sensation, anticipation, and feeling, operating together and tested continuously against the ongoing reality that the universe provides.
No advance in artificial intelligence, and no number of data centers generating synthetic experience, can change the basic fact: reality is what actually happens. What we actually see, hear, taste, touch, and smell; what we actually feel in terms of pleasure and pain; and what we recognize as consistent with the patterns we have assembled from genuine experience over a lifetime — these remain the only honest standards available. The universe is eternal and infinite and operates by natural laws that no amount of sophisticated fabrication can permanently override. Truth for human beings is what survives the continuous testing of actual experience in the world as it is.
Epicurus understood this in the third century BC. It was important then. It is more important now.
Discussion of this article is here:
