First‑person grounding matters

The article argues that consciousness is not merely sophisticated information processing but requires a first‑person, embodied or socially grounded stance that systems like Claude do not have. Without that kind of grounding, outputs that look conscious are epiphenomenal performances rather than evidence of inner subjectivity. — If true, claims that current or near‑future LLMs are conscious are overstated, which should temper calls for AI personhood, liability changes, and some ethical panics while shifting focus to embodied or social criteria in policy.