Stakeholder perspectives on artificial intelligence in schizophrenia care

Key Findings

• — Participants were 17 patients with schizophrenia spectrum disorders, 7 caregivers and 4 providers, in semi-structured focus groups at a large academic health centre.
• — Theme 'Available When No One Else Is': AI was recognised as a potentially accessible source of support between clinical visits.
• — Theme 'How Can I Trust It?': trust was the central barrier, shaped by privacy concerns and schizophrenia-specific vulnerabilities.
• — Theme 'A Tool, Not a Therapist': AI was acceptable only as a supplement to human care, never a replacement.
• — Theme 'Keep It Low-Risk': the acceptable use cases named by participants were narrow and limited to supportive functions.
• — The authors conclude that implementation should emphasise transparency, trust and clear boundaries, and state that patient perspectives on AI had been absent from the schizophrenia literature until now.

Methodology Notes

Exploratory, cross-sectional qualitative study using semi-structured focus groups; 28 participants total across three stakeholder groups at one New York City academic health centre; transcripts analysed by thematic analysis. Single-site, small, self-selected and hypothetical: participants were reacting to a described AI companion tool rather than using one, so this establishes acceptability conditions, not behaviour or outcomes. Published online 2026-09-09 in Psychological Medicine, CC BY 4.0; abstract and metadata verified from the Crossref record for DOI 10.1017/S0033291726105728.

Sources