Artificial Intelligence in Drug Repurposing: Current Advances, Challenges, and Future Perspectives for Precision Medicine
Keywords:
Computational Pharmacology, Drug Repositioning, Machine Learning in Healthcare, Pharmacomics, Algorithmic Bias, Human-in-the-Loop Clinical Decision-Making, Technology Acceptance Survey.Abstract
Background: Traditional drug discovery is constrained by exorbitant costs, extended development timelines, and high clinical attrition rates. Applying artificial intelligence (AI) to drug repurposing offers an efficient strategy to accelerate therapeutic discovery, lower financial barriers, and advance precision medicine. However, stakeholder perspectives regarding its practical integration and challenges remain insufficiently characterised.
Methods: We administered a cross-sectional, questionnaire-based survey to evaluate public and professional awareness, perceived limitations, and ethical considerations surrounding AI-driven drug repurposing. Using a convenience sampling approach, quantitative data were gathered from N=118 respondents representing diverse demographic cross-sections (63.6% female, 34.7% male; 52.5% aged 18–24) and education levels (44.9% undergraduate, 35.6% postgraduate). Descriptive and percentage-based statistical analyses were executed to evaluate familiarity, primary bottlenecks, and governance preferences.
Results: While general AI awareness reached 100%, familiarity with computational drug repurposing was 50.8% n=60). Respondents identified data privacy concerns (26.3%) and algorithm bias (22.9%) as primary computational hurdles. Furthermore, a shortage of skilled interdisciplinary professionals (30.5%) and a deficit in public trust (29.7%) were highlighted as the main adoption bottlenecks. Regarding clinical execution, 44.9% favoured physician decision-making supported by AI systems, while 55.1% projected that AI-guided repurposing will achieve widespread adoption within the next decade.
Conclusion: Although confidence in AI’s transformative impact on precision medicine is high, widespread implementation requires mitigating data privacy risks, building algorithmic transparency, and training cross-disciplinary specialists. Human-in-the-loop clinical frameworks remain essential to ensure safe, ethically aligned adoption.
