The Dual Role of AI in Healthcare: Upskilling the Workforce vs. Risk of Deskilling
Artificial intelligence (AI) is reshaping nearly every facet of healthcare, from diagnostics to education to patient communication. But the direction of this change depends on how AI is applied. On one side, AI has the potential to upskill and cross-skill healthcare professionals, strengthening clinical reasoning and decision-making. On the other, it risks deskilling the workforce when overused as a substitute for human judgment and expertise.
1. AI as a Tool for Upskilling and Cross-Skilling
In medical education, AI is increasingly being used to enhance learning and bridge the gap between knowledge and practice. Adaptive learning systems, virtual simulations, and intelligent tutoring tools allow learners to engage with realistic, case-based problem solving that mimics real-world complexity.
Studies show that AI-integrated education helps learners move beyond rote memorization to develop critical thinking and diagnostic reasoning. For example, BMC Medical Education (2025) mapped how AI is being incorporated into undergraduate medical curricula and found that it can significantly improve competency development when used appropriately. Similarly, Schubert et al. (2024) emphasized that AI education for clinicians is becoming essential for ensuring safe, informed use of emerging technologies.
Beyond education, simulation platforms powered by AI can offer personalized, adaptive training that adjusts in real time to a learner’s decisions, a key advancement over traditional Objective Structured Clinical Examinations (OSCEs). This continuous feedback and flexibility support more effective upskilling and help prepare practitioners for independent clinical reasoning in complex environments.
2. AI as a Driver of Deskilling
While the educational promise of AI is clear, another category of AI applications in healthcare raises concerns. These are systems that aim to replace portions of the healthcare workforce, for example, deploying AI agents to handle first-level and even second-level patient interactions.
While early evidence suggests such systems may reduce workload and improve efficiency, experts warn they could also erode clinical expertise over time. According to Natali et al. (2025), increasing reliance on AI for routine decision-making can inhibit the development and retention of essential clinical reasoning skills. Similarly, research published in Frontiers in Digital Health (2024) notes that automation bias, the tendency to over-trust algorithmic outputs, can lead to a gradual loss of diagnostic intuition and human oversight.
The UCSF Healthforce Center (2025) summarized this dual impact succinctly: AI can either “enhance workforce capability through targeted upskilling or accelerate deskilling when used as a substitution for clinical expertise rather than a complement to it.”
Striking the Right Balance
AI is not inherently good or bad, its impact depends on how it’s implemented. When leveraged as a learning accelerator, it empowers clinicians to practice complex decision-making safely and repeatedly. When used primarily as a labor replacement, it risks hollowing out one of the most cognitively demanding professions in modern society.
The challenge ahead for educators, policymakers, and healthcare leaders is to ensure AI is deployed as a force multiplier for human skill, not as a shortcut that diminishes it.
References
- BMC Medical Education (2025) – Mapping Artificial Intelligence in Undergraduate Medical Education
- Schubert et al. (2024) – AI Education for Clinicians
- Natali et al. (2025) – Artificial Intelligence in Healthcare: Balancing Automation and Human Expertise
- Frontiers in Digital Health (2024) – Automation Bias and Clinical Reasoning Risks
- UCSF Healthforce Center (2025) – AI and Workforce Implications: Annotated Bibliography