READING TEXT
Using AI to talk to the deceased
Read first, then answer the questions. Select an underlined word to see its meaning and save it below.
Original fictional learning scenario. People, dialogue and personal experiences are created for language practice.
In a fictional design workshop, a company proposes a chatbot trained on messages left by a deceased relative. The demonstration is emotionally convincing: familiar expressions appear in replies that the relative never actually wrote. Some participants imagine comfort; others worry that resemblance could be mistaken for continued agency. The distinction matters.
A system can generate language resembling a person's style without knowing what that person would have wanted to say. The group also asks who has permission to supply private messages, how other people mentioned in them are protected and whether users can stop without losing access to the original archive. Nobody resolves every question. They do agree that a moving demonstration is not enough to justify a product. Design choices about consent, disclosure and control must be considered before emotional attachment makes those choices harder to examine.
