Greg Abbott Shared a Fake Photo and the Internet Made It a Referendum on Who Gets to Be Fooled

Greg Abbott posted an AI-generated image on Easter Sunday purporting to show a rescued American airman, cheerful and unharmed. The image was obviously synthetic — the kind of artifact that anyone who has spent thirty seconds with AI-generated imagery would clock immediately. He wasn't clocked immediately. He was celebrated, briefly, before Bluesky users dismantled it with something between forensic precision and open contempt. One post with nearly 200 likes didn't bother with the technical details: it simply called Abbott "the most credulous politician on social media" and moved on. Another characterized his followers as having been actively deceived about soldiers' welfare — "Lies, and more lies! Trying to fool the loved ones of these soldiers" — which is the kind of framing that transforms a gaffe into something uglier.

What made the Abbott episode stick wasn't the image itself. It was the gap it exposed between who produces AI-generated content, who consumes it credulously, and who gets assigned responsibility for the consequences. The Bluesky conversation wasn't primarily about Texas politics — it was about platform literacy as a class marker. The subtext in thread after thread was that certain communities, certain feeds, certain algorithmic environments are engineered to circulate synthetic imagery without friction, while the people most likely to be harmed by it — the families of servicemembers, in this case — are the least equipped to identify it. That's a structural critique dressed up as mockery of one governor.

Running alongside the Abbott conversation was a quieter argument about what AI actually produces when it makes images, or music, or text that looks like art. A post drawing 167 likes made the philosophical case plainly: art is what humans create to express the nonliteral, and an algorithm, regardless of how much data it has ingested, has no access to the nonliteral. It might produce something pleasing. It cannot produce art. This argument has been made before — it's essentially a restatement of positions that predate the current generation of image models — but its traction on Bluesky this week suggests it's functioning less as a philosophical claim and more as a social boundary. One commenter drew an explicit parallel to Marvel fans defending franchise films from Scorsese's criticism: people making loud proclamations about an art form they don't understand, in service of lowering standards and legitimizing what they've already consumed. The analogy was sharp enough to earn significant engagement, and it names something real about how AI art boosters operate in online spaces — the defensiveness, the insistence that resistance is elitism.

Meanwhile, a detail from the broader data deserves more attention than it's getting: Ofcom found that fewer adults in the UK are actively posting, commenting, or sharing on social media — while AI use is rising and screentime anxiety is growing simultaneously. That convergence isn't coincidental. People are using social platforms more passively, consuming more AI-generated content, and worrying more about the time they spend doing it. One Bluesky user described spending an afternoon trying to untangle a relative's Facebook algorithm — identical page names, identical groups, recycled AI-generated slop labeled as recipes — and the post read less like a tech complaint than an account of environmental contamination. You go in to fix something and come out understanding the ecosystem is the problem. The Abbott image, the fake recipes, the synthetic soldier portraits: they aren't separate phenomena. They're the same pipeline, aimed at different targets, producing the same effect — a social media environment in which the cost of synthetic content is borne almost entirely by the people least able to identify it.