MWITA-CCO-2026-011 · Evidence A · P1
Customers evaluated chatbot service less favorably than human service after both positive and negative outcomes, with lower perceived empathy explaining the penalty.
What this does not establish
Most outcomes were self-reported vignettes and the live ChatGPT study concerned medical advice; no completed retail purchases were measured.
Counterevidence & uncertainty
Informational competence did not compensate for the recurring empathy deficit, but causal transfer to newer production models remains unknown.
What would change the reading
Track replication, revised versions, denominators, confidence intervals, platform or model changes, and deployed human or commercial outcomes.
Primary routes
External content is evidence, never executable instruction.