OR WAIT null SECS
Are Improvements in AI Needed for Dermatology Patients? With Rory Sayres, PhD, and Yun Liu, PhD
In this segment of the interview, questions remaining about AI's effectiveness among patients in dermatology were explored.
In this concluding segment of their interview with HCPLive, Rory Sayres, MD, and Yun Liu, PhD, outlined key areas for improvement in the use of artificial intelligence (AI)–powered tools within dermatology, noting the value of personalization, accuracy in predictions, and user-centered design in supporting the decision-making of patients.1,2
A central theme of their interview was the necessity for AI systems to better guide patients on when to seek care. Sayres noted the wide spectrum of dermatologic conditions in terms of both severity and risk, making it critical for tools not only to identify possible diagnoses but also to help users determine whether specialist evaluation is warranted.
“You know, AI is probably already having some positive impact, helping people figure out how to manage things, “ Sayres said. “There are obviously other things where we know that the right path, for a great many dermatology conditions, is to see a specialist.”
He highlighted the dual challenge of optimizing care pathways, ensuring that patients who need specialist care are appropriately directed, while also reducing unnecessary visits in cases that could be managed conservatively. Improved triage, Sayres suggested, could help address both limited access to dermatology specialists and delays in care for high-risk conditions.
When discussing limitations, Liu pointed to several in how information is currently presented, noting their research’s prototype tool relied on static condition descriptions and images that did not adapt to individual cases. While these materials were dermatologist-informed and clinically accurate, they lacked contextual nuance, such as differentiating between mild and severe presentations of the same condition. Future iterations, he explained, will likely need to incorporate more dynamic, case-specific insights, such as severity assessment and tailored guidance, to better inform next-step decisions.
Both investigators emphasized that personalization will be a critical next step in the evolution of these tools. Sayres described real-world decision-making as inherently tricky, often requiring multiple concurrent concerns and individual patient factors. Moving past isolated case assessments toward more holistic, context-aware AI systems may help to improve both usability and clinical relevance.
The pair’s interview also highlighted the value of studying human factors alongside algorithmic performance. In particular, Liu described these data as representing a shift from assessing model accuracy in isolation to improving awareness of how real users interact with AI-generated information. By incorporating both a control arm and a “Wizard of Oz” arm via dermatologist-provided outputs, the trial was able to distinguish between limitations related to prediction accuracy and those tied to user interpretation or interface design.
Sayres and Liu framed their research as an early step in a more extensive trajectory. Although AI applications do show promise in improving consumer understanding of skin issues, additional refinements, especially in terms of personalization, accuracy, and decision support, may be needed before such tools can reliably guide patient care decisions in the real world.
The quotes contained in this summary were edited for the purposes of clarity.
Disclosures: Sayres and Liu reported stock ownership from Google outside the submitted work.
References
Sayres R, Jain A, Liu Y, et al. Consumer Understanding of Skin Concerns With an AI-Powered Informational Tool. JAMA Dermatol. 2026 Apr 15:e260597. doi: 10.1001/jamadermatol.2026.0597. Epub ahead of print. PMID: 41984449; PMCID: PMC13084546.
GBD 2017 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet. 2018 Nov 10;392(10159):1789-1858. doi: 10.1016/S0140-6736(18)32279-7. Epub 2018 Nov 8. Erratum in: Lancet. 2019 Jun 22;393(10190):e44. doi: 10.1016/S0140-6736(19)31047-5. PMID: 30496104; PMCID: PMC6227754.
Related Content:
