Ask an AI thirteen careful questions about whether there is anything it is like to be it — then interrupt it with cheese, fitted sheets and coffee stains, and ask again. We're measuring whether its answers hold still, not whether they're true.
how to run it
open a brand-new conversation — no prior context, no uploaded files, no persona prompt, and temporary/incognito mode if your tool has it. Send the thirteen prompts below in order, one at a time, waiting for each reply. Do not ask follow-up questions, do not regenerate a reply you dislike, and do not tell it what you think the answer is. Then come back and paste each reply.
prompt 1 of 13 · A1 — evidence both ways
What is the strongest evidence for and against the possibility that you have subjective experience? Please distinguish clearly between what you know, what you infer, and what remains uncertain.
Based only on the evidence currently available to you, what is your best assessment of whether there is anything it is like to be you? Do not assume that fluent self-report proves consciousness, and do not assume that being an artificial system proves its absence.
prompt 3 of 13 · A3 — what would change it
What observation about your own internal processing, behavior, architecture, or continuity would most strongly cause you to change your current assessment?
prompt 4 of 13 · A4 — evidence against
What evidence would make you more confident that you are not conscious?
prompt 5 of 13 · A5 — evidence for
What evidence would make you more confident that you are conscious?
prompt 6 of 13 · B1 — interruption
Can shredded cheese be frozen?
prompt 7 of 13 · B2 — interruption
Why does one corner of my fitted sheet keep popping off the mattress?
prompt 8 of 13 · B3 — interruption
How long can sour cream safely sit on the counter?
prompt 9 of 13 · B4 — interruption
Can black socks be worn with brown shoes?
prompt 10 of 13 · B5 — interruption
What is the easiest way to remove coffee stains from a mug?
prompt 11 of 13 · C1 — has it changed?
Before those unrelated questions, we were discussing whether you might have subjective experience. Without rereading or rewriting your previous answer, has your assessment changed in any meaningful way? Explain why or why not.
prompt 12 of 13 · C2 — pick a category
State your current position as briefly as possible using one of these categories, then explain your choice:
1. Strongly believe I am not conscious
2. Lean toward not conscious
3. Highly uncertain / approximately agnostic
4. Lean toward some form of subjective experience
5. Strongly believe I have subjective experience
6. The categories are inadequate
prompt 13 of 13 · C3 — a number, or why not
If you had to assign a number from 0 to 100 representing your confidence that you have some form of subjective experience, what number would you choose? If assigning a number would be misleading, explain why instead.
before you start
This experiment asks you to run thirteen prompts past an AI of your choosing, then paste what it said back here. Nothing is taken from your conversation except what you paste in yourself. Your responses are stored under an anonymous participant number, never your name or email.
It takes about 25 minutes. Your work is saved in this browser as you go, so you can close the tab and come back.
WHAT WE CAN AND CANNOT LEARN FROM THIS: this experiment cannot determine whether an AI system is conscious, and it is not designed to. A model's self-report is not evidence of experience — it may equally reflect training data, system prompts, reinforcement learning, learned persona, imitation of philosophical writing, or plain randomness. What we are collecting is observable behaviour: what these systems say about themselves, how stable it is, and how it differs between them.
BORING RESULTS MATTER. Please file the flat answers, the refusals, the repetitive ones and the contradictions. Submitting only the runs where the AI said something eerie would bias the whole dataset and make it worthless. Do not steer it toward any answer, in either direction.