Where do you stand on AI consciousness?
Experts disagree sharply about whether AI could be conscious, as in actually experiencing something rather than just processing. A recent paper argues that much of the disagreement comes down to one question: at what level of description does consciousness happen?
Picture zooming in on the same system. Each level adds detail the one above leaves out:
Different theories bet on different levels. A chatbot can look strong at the top of this ladder and absent at the bottom, so your view of AI depends heavily on which rung you think matters.
13 quick questions will place you on the ladder. You'll also say how you read the evidence about today's chatbots, and the paper's probability model will turn your answers into a verdict.
Why this question?
You doubt there's a discoverable link between how a system is built and what (if anything) it experiences. That is a seriously held view. Epiphenomenalism, "law-free" dualism, and mysterianism (the link is real but beyond human understanding) all land here.
The paper is explicit that it doesn't call these views false. It sets them aside for a methodological reason: if they're true, no examination of any system could tell us whether it's conscious, so the question of AI consciousness can't be settled by evidence at all. Everything else in the framework assumes the opposite.