The use of artificial intelligence in diagnosing emergency cases: a crosssectional study of physicians’ perspectives on the effectiveness of AI systems in improving diagnostic accuracy and reducing time to diagnosis

Abstract

Background: Artificial intelligence (AI) has the potential to improve diagnosis accuracy and speed in emergency rooms; yet, its acceptance and use by emergency physicians in Saudi Arabia remain limited.Methods: A descriptive cross-sectional survey was issued to emergency physicians in Saudi Arabia. The questionnaire looked at demographics, AI exposure, perceived benefits and problems, and overall views toward AI. The data were evaluated with descriptive statistics, chi-square tests, ANOVA, and logistic regression.Results: A total of 352 physicians participated, the majority of whom were young residents. Only 11.6% said they used AI on a regular basis. The majority of physicians (65.3%) had a negative attitude, whereas 6.8% were optimistic. Consultants had more favorable perceptions than residents. Attitudes differed considerably by region, age, gender, country, and professional level (p < 0.05). Key perceived benefits included increased diagnostic accuracy and speedier decision-making, while major hurdles included a lack of training, system inaccuracy, and difficulties interpreting AI outputs. Male and non-Saudi physicians had higher positive sentiments.Conclusion: Despite acknowledging AI›s potential, practical adoption in emergency contexts remains low due to poor training and dependability issues. Structured AI training and system development are required to ensure a safe and effective integration into emergency practice.