Design guidelines for the therapeutic use of AI in grief

Abstract

Within grief, chatbots are being used for therapy and as “griefbots”, impersonating the dead. Users risk data breaches and psychological attachments to monetised services. To determine human-centred guidelines for the responsible implementation of grief AI, a qualitative study was conducted, consisting of therapist interviews (N = 4), a survey of individuals who experienced grief (N = 49), and interviews with a subset of respondents (N = 4).

Participants expressed low trust in the efficacy and data security of therapeutic chatbots and strongly rejected griefbots, citing ethical and psychological concerns. Therapists also highlighted associated risks, but identified opportunities for their use as complementary therapy tools, emphasizing a need for professional supervision and usage restrictions.

Guidelines include personalisation, data privacy and security, professional oversight, and restrictions by age and psychological characteristics. The novel RUDA framework (Responsible Usage, Development, and Administration) is proposed, mitigating associated risks by outlining accountability by actor for the responsible implementation of grief AI.