The detachment of responsibility: how and when AI-powered digital avatars undermine behavioral intentions in cancer screening persuasion

In plain English

In a 2x2 experiment with 211 Chinese participants, a pancreatic cancer screening message delivered by an avatar replica of a real person produced lower screening intentions than the identical message from the human original. The mechanism was perceived communicative responsibility: audiences held the avatar less accountable for what it said, and that penalty was larger when the speaker was a doctor than when it was a blogger. The practitioner implication is uncomfortable for the clone-your-expert business model — the authority you are borrowing may not survive the transfer, and it degrades most where expertise matters most.