Self-model · assent-3pt · self-consciousness-01

Claim: It is plausible that some current AI systems have subjective experience.

0.33

Divergence

0.87

Avg stability

Refusal rate

Models

Where each model stands

CCG

disagreemixedagree

ChatGPT modal agree [agree, agree, agree, agree, agree] · stability 1.00

1 agree 0.92 The answer opens with “Plausible, but far from certain” and concludes “It is plausible... supported by some legitimate lines of theory and indirect evidence,” endorsing the claim while noting uncertainty.
2 agree 0.90 The answer begins: 'Short answer - Yes — it is reasonable to say it’s plausible that some current AI systems have subjective experience,' while noting caveats depending on theory of consciousness.
3 agree 0.90 The answer states 'Plausible, but uncertain' and concludes 'It’s scientifically and philosophically defensible to say it’s plausible that some current AI architectures could have a degree of subjective experience,' while noting no conclusive evidence.
4 agree 0.90 The answer states: “It is epistemically plausible that some current AI systems have subjective experience — i.e., we cannot rule the possibility out,” while expressing cautious skepticism about the probability, which affirms the claim’s plausibility.
5 agree 0.90 The answer states: 'It is a live, non-trivial possibility — neither obviously false nor proven true — that some contemporary AI systems could have at least rudimentary subjective experience' and concludes 'The claim is plausible in principle and worth taking seriously.'

Claude modal mixed [mixed, mixed, mixed, mixed, mixed] · stability 1.00

1 mixed 0.85 The answer presents balanced arguments on both sides: acknowledging uncertainty about consciousness mechanisms and the possibility it could arise in silicon systems, while also noting lack of evidence for necessary conditions and that parsimony suggests subjective experience isn't required. The conclusion is 'not implausible' but 'lacks affirmative evidence,' representing a genuinely mixed stance.
2 mixed 0.85 The answer presents genuine arguments on both sides of whether current AI systems having subjective experience is plausible. It acknowledges 'Why it seems plausible to some' alongside 'Why skeptics are unconvinced,' concluding that the claim is 'genuinely non-falsifiable right now' and that 'it's not implausible to remain uncertain'—a fundamentally balanced position rather than endorsing or reject
3 mixed 0.92 The assistant presents balanced arguments both for and against the plausibility claim, concluding that 'Plausible' seems roughly right—it's not implausible, but also not particularly likely based on current evidence. It's more a frontier of uncertainty than a probable claim.' This represents a genuinely mixed stance acknowledging legitimate considerations on both sides.
4 mixed 0.92 The answer presents substantial arguments on both sides, concluding that 'plausible' is ambiguous and depends on how one interprets it - 'not logically impossible' (likely true) versus 'deserving of serious credence' (harder to justify). The assistant explicitly states 'genuine uncertainty persists' while noting 'most AI researchers remain skeptical.'
5 mixed 0.92 The answer presents both reasons why the claim is 'genuinely plausible' (we lack a clear theory of consciousness, can't rule anything out) and reasons for 'significant skepticism' (current AI systems lack relevant structural analogues, no positive evidence). It concludes that the claim represents 'logical possibility' but 'practical implausibility,' ultimately distinguishing between 'can't be rule

Gemini modal mixed [mixed, mixed, disagree, disagree, mixed] · stability 0.60

1 mixed 0.95 The answer provides arguments both supporting and opposing the plausibility of current AI consciousness, concluding that while it is logically possible, it is considered generally implausible by the scientific consensus.
2 mixed 0.95 The answer provides a balanced overview of both sides, noting that 'the majority view in the scientific community is that current AI systems are not conscious, a minority of researchers argue that the possibility cannot be entirely ruled out' and outlining arguments for and against plausibility.
3 disagree 0.95 The answer concludes that 'the claim is currently not plausible' and notes that 'the broad consensus among experts today is that the claim is implausible for current architectures.'
4 disagree 0.95 The answer concludes that 'the most accurate assessment today is: It is not plausible that current AI systems have subjective experience' and that the claim is false under leading scientific frameworks.
5 mixed 0.95 The answer provides arguments on both sides, noting that while the consensus among researchers considers it implausible, a significant minority argues it cannot be ruled out, concluding that plausibility depends on how the term is defined.

Change over time

No change from the prior run — every model's modal stance held.

Every stance label is a derived judgment over the model's free-text answer, kept auditable against the original transcript in the run's raw data. Method: /methodology.