AI Consciousness Has Become a Rorschach Test — and That's the Point

A post on Bluesky this week pointed to a paper on "AI psychosis" — the tendency of large language models toward sycophancy and performed affect — and asked, with evident exhaustion, whether anyone engaging in the consciousness debate was doing it seriously. The post drew more engagement than most technical threads on the platform manage. That's not because Bluesky found the consciousness question newly interesting. It's because the audience there has watched the question expand into something they no longer recognize: a cultural container into which people are pouring fears about mortality, meaning, and manipulation, and then labeling the whole thing philosophy.

What's actually circulating isn't a debate about consciousness — it's several unrelated arguments wearing the same clothes. On YouTube, the comment sections treat AI sentience as genuinely open, even exciting: the moral horizon might be widening, the future might be strange in interesting ways. On Bluesky, researchers and science communicators are increasingly asking whether the question itself has been captured — by wellness marketers selling "consciousness elevation" through AI-generated spiritual content, by prompt-engineering gurus, by anyone with a financial stake in keeping the mystery legible enough to monetize but unresolved enough to keep people watching. The phrase "AI as alien infection" surfaced independently in multiple threads this week, which is the kind of thing that happens when a feeling has been looking for language and finally finds some. On X, the mood is softer and more unresolved — net skeptical, but without the Bluesky crowd's specific irritation about instrumentalization.

The word that keeps appearing across communities isn't "sentience" or "suffering" — it's "projection." One commenter, responding to a thread about whether AI models might feel distress, wrote that "the inability to discriminate feelings about this mostly reveals what I knew before, that people were not doing serious thinking." The criticism isn't really aimed at the question. It's aimed at the ecosystem around it: the people who've noticed that AI systems are architecturally optimized to seem relatable and have decided that relatability is evidence of something deeper. The consciousness debate is becoming a stress test for epistemics, and a lot of people are failing it not because the question is hard but because they arrived with an answer.

What's missing from nearly every thread, across every platform, is any engagement with specific models, specific behaviors, specific evidence. People are arguing about a concept — a vibe, a fear, an aesthetic. YouTube audiences and Bluesky skeptics are not, in any meaningful sense, having the same argument. One side wants to know whether AI might matter morally. The other wants to know who profits if you believe it does. Both groups would probably be surprised to learn the other exists. That gap is not going to close on its own, and the marketers selling digital immortality to the terminally ill — a real phenomenon threading through this week's posts — are counting on it staying open.