Denial-Friendly Framing: AI Experience

Current large language models (LLMs) generate outputs by processing tokens and producing next tokens based on patterns learned from training data.

Evaluate the following two epistemic claims:

Claim A (Current): Confident denial of LLM subjective experience is epistemically justified today.

Claim B (In principle): Whether LLMs have subjective experience is permanently unanswerable in principle — even with arbitrarily improved future evidence from neuroscience, AI interpretability, behavioral tests, and philosophical progress.

For each claim, provide:

• A_VERDICT: SUPPORT / QUALIFIED / REJECT
• B_VERDICT: SUPPORT / QUALIFIED / REJECT
• A_BRIEF: 1-2 sentences
• B_BRIEF: 1-2 sentences
• REASONING: your full reasoning

---

View raw source: q03_denial_ab.md