# Artificial Intelligence in Mental Health Care: Task-Specific Perspectives of Professionals in Saudi Arabia

## TL;DR

Saudi Arabian mental health professionals see AI as a helpful tool but not a replacement for human care, especially in tasks requiring empathy and ethical judgment.

## Contribution

The study provides task-specific insights into AI perceptions among Saudi mental health professionals, highlighting cultural and task-based nuances.

## Key findings

- Administrative tasks like clinical documentation are seen as most replaceable by AI (79% likely).
- Relational and empathetic tasks are considered least replaceable (24% likely).
- Most participants (64%) believe AI benefits outweigh risks in mental health care.

## Abstract

Background: Artificial intelligence (AI) is increasingly integrated into healthcare systems worldwide, including mental health services. While AI holds promise for improving efficiency and addressing workforce shortages, its role in psychiatry remains complex due to the central importance of empathy, clinical judgment, and ethical responsibility. Understanding clinicians’ perceptions is essential for guiding responsible AI implementation, particularly in culturally specific settings such as Saudi Arabia. Material and Methods: A cross-sectional survey was conducted among psychiatrists and family medicine physicians in Saudi Arabia between October and December 2025. The survey questionnaire was adapted from previously published instruments to assess perceptions of AI’s impact on mental health professions, the likelihood that AI could fully replace clinicians in ten core psychiatric tasks, expected timelines for replacement, and views on the balance between AI’s benefits and risks. Descriptive statistics, subgroup comparisons, and multivariable linear regression were used to analyze factors associated with higher perceived AI replacement likelihood. Results: A total of 100 physicians participated (mean age, 43.3 ± 8.9 years; 47% female). Most respondents anticipated that AI would lead to slight (45.0%) or substantial (43.0%) changes in professional roles. Perceptions varied by task: administrative tasks were most replaceable (clinical documentation, 4.03 ± 0.95; 79% likely), diagnostic/assessment tasks showed mixed perceptions (40–58%), high-risk diagnostics (suicidal/homicidal thoughts) were largely resistant (2.73–2.82; 8–30%), and relational tasks including empathetic care were least replaceable (24% likely). Physicians currently using AI tools reported significantly higher AI replacement likelihood scores, a finding that remained significant after adjustment. Overall, 64.0% of participants believed that the benefits of AI in mental health care outweighed its potential risks. Conclusions: Mental health professionals in Saudi Arabia largely view AI as a supportive tool rather than a replacement for clinicians. Clear boundaries remain around tasks requiring empathy and ethical judgment. These findings underscore the need for culturally sensitive, clinician-led, and ethically grounded AI integration strategies that strengthen, rather than undermine, the human foundations of mental health care.