Medical students’ perceptions of artificial intelligence following the “Robomedic” AI-awareness campaign: a pre–post survey study in Saudi Arabia

Background Although artificial intelligence (AI) is increasingly shaping healthcare and medical education, evidence remains limited on how brief AI-awareness activities influence medical students’ perceptions, self-reported understanding, and concerns about future professional roles. Understanding these short-term changes is important for evidence-based curriculum design in the AI era. Using the Technology Acceptance Model and the Unified Theory of Acceptance and Use of Technology as interpretive lenses, this study examined short-term changes in medical students’ perceptions and self-reported understanding of AI following Robomedic, a student-led AI-in-medicine awareness campaign. Methods A single-cohort pre–post survey was conducted at King Faisal University College of Medicine, Saudi Arabia. The baseline survey was administered three days before Robomedic, and the post-campaign survey was administered within one week after the event. Robomedic was a brief interactive campaign that included expert-led talks and educational booths addressing AI applications, benefits, limitations, ethical considerations, and the role of AI as a supportive rather than replacement tool in clinical practice. Of 91 students who completed the baseline survey, 50 completed both surveys and were matched using a four-digit academic identification code. The questionnaire included 15 Likert-scale items grouped into perceived-benefit and concern-related domains. The composite AI perception score was calculated as the mean of all 15 items after reverse coding concern-oriented items, with higher scores indicating more positive perceptions. The primary outcome was change in the composite score, assessed using a paired-sample t-test. Results Among the matched participants ( N = 50), the findings showed a mixed pattern: perceived benefits, self-reported understanding, and confidence increased, while replacement-related concerns also became more prominent. Agreement that AI would play an important role in healthcare increased from 41/50 (82%) to 48/50 (96%). Agreement with understanding basic AI principles increased from 18/50 (36%) to 46/50 (92%), comfort with AI terminology from 23/50 (46%) to 43/50 (86%), and confidence in using basic healthcare AI tools from 29/50 (58%) to 45/50 (90%). The composite perception score increased by 0.28 points on the 1–5 scale (95% CI, 0.16–0.40), representing a moderate short-term perceptual shift, t(49) = − 4.52, p < .001; Cohen’s d = 0.64. At the same time, replacement-related concerns increased. Agreement that some specialties may be replaced by AI rose from 17/50 (34%) to 25/50 (50%), agreement that AI may replace radiologists rose from 14/50 (28%) to 22/50 (44%), and reluctance to consider radiology as a career increased from 14/50 (28%) to 25/50 (50%). Thus, the post-campaign findings reflected increased perceived benefit and self-reported understanding alongside increased concern, rather than a uniformly favourable perceptual change. Conclusion This study found a mixed pattern of short-term perceptual change following Robomedic: medical students reported greater self-reported understanding, confidence, and perceived relevance of AI, while replacement-related concerns also increased. These findings suggest that brief AI-awareness activities may benefit from explicitly addressing AI limitations, automation anxiety, specialty-replacement narratives, and the continuing role of human clinical judgment, for example through structured discussions that distinguish task automation from physician replacement.