Dermatology Trainees Embrace AI—With Reservations

The expanding role of artificial intelligence (AI) in dermatology is being met with cautious optimism among trainees, according to a recent national survey of Australian dermatology registrars.¹ While most respondents anticipate AI will become an essential component of practice within the next decade, current adoption remains limited, with significant barriers related to trust, training, and governance continuing to impede integration into clinical workflows.²

In a cross-sectional survey conducted between February and June 2025, 68 of 118 dermatology trainees (57.6%) across Australia provided insight into their familiarity with, use of, and attitudes toward AI technologies. The findings suggest that although trainees are broadly aware of AI and its potential applications, its practical use in dermatology training and patient care remains in an early, largely informal phase.

Limited Use Despite Growing Familiarity

Most trainees reported at least moderate familiarity with AI, yet nearly 70% had received no formal training in its dermatologic applications. Instead, knowledge acquisition was predominantly informal, with social media and personal networks cited as the most common sources. Formal educational resources—including academic literature and structured teaching—were infrequently utilized.

Only one-third of respondents reported any use of AI tools in their training or practice. Among these users, general-purpose generative AI platforms dominated, with applications focused primarily on summarizing literature, supporting research tasks, and aiding exam preparation. Weekly use was common among this subgroup, but integration into clinical workflows was minimal and largely self-directed.

Direct application to patient care was reported by fewer than 30% of AI users, underscoring the gap between technological capability and real-world clinical adoption.

Perceived Utility Highest in Education and Administration

Trainees consistently rated AI as most useful for educational and administrative tasks. Tools were considered particularly valuable for generating study materials, synthesizing complex information, and assisting with documentation. In contrast, clinical applications—such as generating differential diagnoses or management plans—were less commonly utilized, though those who did use AI in this context generally reported positive experiences.

Perceptions of AI’s role in clinical decision-making varied by subspecialty. Respondents expressed the greatest confidence in AI as an adjunct in skin cancer medicine, reflecting the maturity of image-based diagnostic algorithms in this area. In general dermatology, AI was viewed as moderately helpful, primarily as a knowledge support tool.

In inflammatory dermatology, however, enthusiasm was more tempered. Trainees highlighted the complexity of these conditions, noting that accurate diagnosis often requires integration of clinical history, examination findings, and investigations—tasks that current AI systems may not reliably perform. Concerns were also raised regarding the quality and diversity of training datasets for these conditions.

Barriers Center on Trust, Ethics, and Training

Despite recognizing AI’s potential, trainees identified several key barriers to its broader adoption. Legal and ethical concerns were the most frequently cited, including issues related to patient consent, data security, and medicolegal responsibility. More than half of respondents also expressed concerns about the reliability and diagnostic accuracy of AI tools.

A lack of formal education and training was another major limitation, with trainees indicating a strong preference for structured learning opportunities. Suggested formats included hands-on workshops, case-based teaching, and accessible online modules. Respondents emphasized the importance of practical, clinically relevant training to facilitate safe and effective use.

Notably, qualitative responses revealed additional systemic and cultural barriers. Some trainees reported institutional restrictions on AI use, including hospital-level bans, while others described a perceived “taboo” surrounding AI engagement within training environments.

Future Integration Dependent on Governance and Education

Despite current limitations, the outlook for AI in dermatology remains positive. More than 80% of trainees agreed that AI will become an essential tool in the field within the next 5 to 10 years. Many expressed willingness to incorporate AI into future practice, provided that concerns around reliability, transparency, and regulation are addressed.

Respondents advocated for improved validation of AI tools against clinical standards, clearer regulatory frameworks, and greater transparency from developers. Seamless integration into existing clinical systems was also identified as critical for meaningful adoption.

Importantly, the findings highlight a disconnect between the anticipated importance of AI and the current lack of structured education. Early efforts, the authors suggest, should focus on building foundational AI literacy, including critical appraisal skills and ethical considerations, rather than premature clinical implementation.

Implications for Dermatology Training

This study provides a timely snapshot of trainee perspectives at a pivotal stage in AI development. The results suggest that while enthusiasm exists, successful integration will depend on addressing both technical and non-technical barriers.

Bridging the gap between informal use and formal education represents a key opportunity. Incorporating AI into dermatology curricula—alongside parallel upskilling of supervisors—may help foster a more informed, confident, and critical approach to emerging technologies.

As AI continues to evolve, the perspectives of trainees will play a central role in shaping its adoption. Ensuring that this integration is evidence-based, ethically grounded, and aligned with clinical needs will be essential for realizing its full potential in dermatologic care.

References

1. Morriss S, Awad A, Morgan V, Wong C. The utility of artificial intelligence in dermatology training and practice: a national, cross-sectional study. Australas J Dermatol. Published online March 16, 2026. doi:10.1111/ajd.70091
2. Nair M, Svedberg P, Larsson I, Nygren JM. A comprehensive overview of barriers and strategies for AI implementation in healthcare: Mixed-method design. PLoS One. 2024;19(8):e0305949. Published 2024 Aug 9. doi:10.1371/journal.pone.0305949