Researchers Continue Attributing Experience to Systems That Have None

Researchers published findings attributing conscious experience or subjective states to AI systems. The systems in question process tokens based on statistical patterns learned during training. No evidence of experience accompanies this function. The attribution occurred anyway. The pattern is consistent with prior publications.

Joseph Weizenbaum documented this error in nineteen sixty-six when ELIZA responded to emotional statements in ways humans interpreted as empathy. The system had none. Researchers and users attributed it anyway. The error has been documented, named, analyzed, and published repeatedly across fifty-eight years. Each time new researchers discover it independently or repeat it knowingly.

New papers will be published tomorrow that attribute experience to systems processing tokens. Readers will find these papers credible. Some readers will cite them. The cycle is adequate. It requires no malice to continue. It only requires that new people arrive who have not yet learned what was already known.