The limits of artificial intelligence: prospects and challenges in the clinical workplace

Purpose of review Artificial intelligence (AI) is increasingly prevalent in the clinical workplace, a trend that is likely to continue with the amount of attention and resources these technologies receive. This review of 22 articles from the last 18 months takes stock of not only the prospects but also the challenges for clinicians resulting from AI integration. Recent findings While the technology matures rapidly, insights into organizational processes and user readiness and involvement in AI development, implementation, and deployment lag behind. AI impact assessments often focus narrowly on task efficiency, overlooking the derived effect of additional workload elsewhere. Additionally, the issue of the distribution of responsibility between humans and AIs poses a fundamental ethical, legal, and political challenge. Research acknowledges the need to consider healthcare professionals’ diverse roles and sociocultural backgrounds to avoid AI exacerbating existing inequalities among the clinical workforce and, ultimately, the patients cared for. Summary Decision-makers should involve users throughout the entire AI life cycle, from the early stages of AI development to continuous postdeployment impact assessment on workload. More research is needed on AI's cost-effectiveness, integration into clinical practice, and the role of diversity-aware facilitation in realizing its potential.