Jean Baudrillard’s Simulacra and Simulation explores the uneasy cohabitation of reality, symbols, and society – particularly how images and media shape our perception of what is real. What we deem to be true and what is true begin to drift apart, and with that drift comes a shift in how we behave when confronted with the difference. In 1981, Baudrillard outlines four phases of the image. Phase One is an image that reflects reality: a photograph of your dog, taken by you. Phase Two is an image that distorts reality: an advertisement where the burger looks twice as big as the real one, glistening under the lights, more delicious than it will ever taste. Phase Three pretends to reflect a reality that is thinning out, dissolving: an influencer’s Instagram feed presenting a life that does not flow as lived – strategically cropped images, moments captured only because they are meant to be captured and posted, a performance standing in for a life assumed to exist somewhere off-screen. Phase Four is where even that assumption dissolves, where reality is no longer taken into consideration at all – it begins to collapse. A brand selling “authenticity” as a concept, referencing a heritage that no longer exists; political aesthetics detached from policy; a social media persona branding themselves as radical without being able to speak to any radical positions, voting conservatively, or simply not engaging in politics in real life.
Baudrillard proposed his theory as an autopsy of reality, a way to trace how reality learned to disappear in front of our eyes, through our own actions. In the third phase, the image claims to represent reality in its unaltered state, yet the original is receding. What we’re left with is a performance of truth, speckles of it may be dusted throughout, but it’s miles away from its true form. In the fourth phase, the image no longer refers to anything at all, because it simply exists without the crutch of reality. One would expect outrage when coming to terms with the fourth phase, yet there is only quiet resolve, there is comfort in the absence of reality it seems. We stopped concerning ourselves with whether something is real, but we simply ask if it can be shared, if profit can be made on the back of it. The irony, of course, is that this condition feels uncannily familiar, even intimate. We participate in it daily, earnestly, nostalgically, while insisting – often loudly – that we are still grounded, still authentic, still in touch with something original. We do not mourn the loss of reality so much as we forget to look for it, reassured by systems that function perfectly well without it. Baudrillard would likely smile at this insistence. It is, after all, one of the system’s more elegant effects.
What Baudrillard didn’t predict was generative AI, not because it contradicts what he theorised about – but it completes it too well. It’s simply too good. In the fourth phase, AI doesn’t distort reality, it bypasses it completely. Where earlier images still required an original, however distant, quiet, or fictional, AI-generated images no longer rely on lived experience, memory, or even intention. Generative AI enables optimised meaningless production. Meaning is no longer authored – it is assembled by a robot. In this sense, the image doesn’t claim to tell a truth, it doesn’t even claim to lie, it just exists and you don’t even have to feel anything about it. You don’t even matter anymore. AI-generated images don’t require context, they appear with no prior, endlessly reproducible. Where a photograph was once proof that something had been there, proof of a reality that either currently exists, once existed, or exists in a different form, AI produces images that don’t need reality to exist. Reality is irrelevant to their existence and production.
What is perhaps most unsettling is how little resistance this has been met with so far. Instead, there is ease, comfort in its producability. Images arrive fully formed, convincing enough, familiar enough, and we accept them not because we believe in them, but because belief isn’t really a requirement anymore. We scroll past, we save, we share, we move on. We’re more concerned with whether an image creates engagement instead of whether it’s real. Does it fit the narrative? If yes, then it doesn’t have to reference anything anymore. Baudrillard warned that the real would disappear behind the simulacrum. AI suggests something colder still: representation continues perfectly well without needing someone to represent. The image no longer answers to a world outside itself, and no one seems particularly troubled by this. If there is a loss here, it is not only of reality, but of authorship, of responsibility, of the expectation that images should carry the burden of truth.