Particle.news
Download on the App Store

Psychologists’ AI Use Jumps Even as Safety Concerns Rise

Recent research supports supervised, interim use for mild symptoms, not full clinical care.

Overview

  • An APA survey of 1,700 psychologists found 56% used AI tools in the past year, up from 29% in 2024, with 29% using them monthly.
  • Use remains largely administrative, led by email drafting, note and article summarization, content creation, and documentation; only 8% reported diagnostic support and 5% patient-facing chatbots.
  • Concerns are widespread, with 92% voicing worries, including data privacy (67%), societal harms (64%), algorithmic bias (63%), and hallucinations (60%), all higher than last year.
  • A separate report found 52.4% of clients already turn to general chatbots for support between sessions and preparation, yet most fear faulty advice (82.4%) and data risks (73%) and want AI to complement therapy.
  • Controlled studies show a therapist-trained LLM can reduce depression, anxiety, and eating-disorder symptoms over eight weeks, but other research flags unsafe and stigmatizing responses to psychosis and suicidality in general-purpose models.