APA Looking Beyond Webinar - AI and the Future of Psychiatric Assessment: A Focus on Equity-View Presentation

Video Transcription

Video Summary

The APA “Looking Beyond” webinar on <strong>AI and the future of psychiatric assessment with an equity focus</strong> explored how AI could both improve care and amplify historic harms. Moderator <strong>Dr. Stephen Chan</strong> framed AI’s rapid move into clinical practice as creating opportunities (earlier detection, consistency, personalization) alongside serious risks if bias is automated at scale.

<strong>Karis Myrick</strong> (Inseparable; disability rights and lived-experience leader) warned that AI models trained on biased clinical data can reproduce over-, under-, and misdiagnosis for Black, Indigenous, Latine, LGBTQ+, disabled, rural, and low-income communities—often with a misleading “objective” veneer. She emphasized that AI can miss context (racism, poverty, trauma) and urged clinicians to watch for overreliance on AI, demographic score skews, poor transparency about training/validation, limited representation in datasets, and patient distrust or feeling unheard. Consent, data ownership, privacy, and today’s surveillance climate were central concerns, especially for video recording and digital phenotyping.

<strong>Dr. Peter Yellowlees</strong> (UC Davis emeritus; telepsychiatry pioneer; Async Health co-founder) agreed and stressed AI’s cultural limitations, sharing cross-cultural examples where beliefs could be misread as psychosis without cultural interpretation. He advocated <strong>“hybrid care”</strong>—AI as a tool that augments but never replaces clinicians. He described AI-assisted, clinician-designed avatar interviews and argued that responsibly stored patient video could strengthen documentation and future algorithms, while acknowledging patient control and consent.

Panelists discussed regulation (state-level safeguards where federal action is limited), workforce fears (AI should help treat more unmet need, not replace clinicians), and the need for <strong>digital literacy and informatics training</strong> across clinical education, with lived experience meaningfully included in design and oversight.

Keywords

AI in psychiatric assessment

health equity and bias

algorithmic bias in clinical data

digital phenotyping and surveillance

informed consent and data ownership

privacy of patient video recordings

hybrid care (AI-augmented clinicians)

cross-cultural psychiatry and misdiagnosis

telepsychiatry and avatar interviews

digital literacy and informatics training