Explore the scholarly record.

Abstract

Abstract Background: The integration of artificial intelligence (AI) into medical research and education is rapidly advancing. Despite being aware, Nigerian medical students and doctors demonstrate low utilization of AI in medical research with a major barrier being limited access to structured training in its research applications. This study aims to assess and measure the impact of a well-structured intervention on the perception, knowledge and ethical concerns of AI in medical research among medical students and doctors in Nigeria. Methods: A pre-test posttest quasi-experimental design was utilized among medical students and doctors across Nigeria. The intervention was delivered as a 2-day online webinar workshop via Microsoft Teams and data were collected via questionnaires consisting of Likert-scale questions assessing demographic information, baseline perceptions across eight domains derived from the Health Belief Model supplemented with additional constructs: perceived knowledge and ethical concerns, and changes following the intervention. Data were analyzed with R version 4.5.2 and IBM SPSS Statistics Version 27. Statistical significance was set at p-value < 0.05. Results: 135 participants were included in the study. 77.0% of participants had no formal AI training, though 31.9% had attended a prior AI workshop. At baseline, perceived knowledge of AI was the lowest while perceived benefits was the highest. Following the training intervention, statistically significant improvements were observed across seven of the eight domains. Knowledge increased from 3.16 ± 0.73 to 4.18 ± 0.50 (Z = − 9.787, p < 0.001, r = 0.84), Perceived susceptibility from 3.83 ± 0.72 to 4.18 ± 0.72 (Z = − 5.022, p < 0.001, r = 0.43), and Perceived severity from 3.75 ± 0.75 to 4.04 ± 0.72 (Z = − 4.540, p < 0.001, r = 0.39). Perceived benefits showed a smaller but significant improvement 4.17 ± 0.64 to 4.33 ± 0.53 (Z = − 2.654, p = 0.008, r = 0.23), while Perceived barriers decreased significantly (Z = − 5.147, p < 0.001, r = 0.44). Cues to action and Self-efficacy also improved significantly (p < 0.001 for both; r = 0.53 for each). Ethical concerns showed no significant change (Z = − 0.559, p = 0.576). Satisfaction with the intervention was high, with 93–97% of participants endorsing the training as relevant, well-organized, and worth recommending to colleagues. Conclusion: Perception of AI among Nigeria medical students and doctors is good, however research-oriented literacy and application is limited. Structured intervention can significantly improve the perception and knowledge of AI in medical research, however ethical concerns about AI in medical research need to be given more emphasis in future interventions.