Answers are grounded in this paper and cite their sources.
Background: Artificial intelligence (AI) is transforming healthcare by improving care, efficiency, and cost benefit, yet limited research compares its benefits across individual, organizational, and system levels to guide effective integration. Aims: This review examines how the adoption of AI can benefit healthcare by improving individual outcomes, enhancing organizational efficiency, and strengthening the healthcare system. Methods: Two reviewers screened studies published up to December 2023 from four databases-a three-dimensional framework identified AI benefits for individuals, organizations, and the broader health sector. Results: A total of 92 articles met the inclusion criteria, showing AI provides meaningful benefits across all dimensions. The individual benefits dimension is precise but often context-specific, showing substantial gains in early detection, diagnosis, and monitoring. AI reduces the workload by 19%-50%, improves reading times by 21%-54%, and enhances cancer detection by 20% and treatment timing. The organizational benefits dimension encompasses AI-optimized workflows and cost reductions, enhancing efficiency in clinical settings and workflow management. The health sector system dimension shows the most substantial evidence, highlighting improved patient outcomes, workflow efficiency, data access, training, and overall system advancement. Despite these advantages, challenges remain regarding data integrity, patient safety, and privacy under strict healthcare regulations. Conclusion: AI offers benefits at all levels, with the most substantial evidence being at the system level, demonstrating clear improvements in patient care and workflow. Individual and organizational gains are promising, though more validation is needed. Future adoption should prioritize cost-benefit analysis and system-wide AI integration while evaluating impacts at all levels.
No takes yet. Share an insight, caveat, or question.
Gebremeskel et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: