Research into how AI can help users understand skin conditions

Details

Google Research shared results from two papers published this week in JAMA Dermatology examining how AI tools affect consumer understanding of skin conditions. The work builds on years of Google's dermatology AI foundation, including differential diagnosis models and the SCIN dataset. The larger quantitative study showed 2,345 survey participants retrospective de-identified skin cases and found that structured AI assistance improved participants' ability to name the condition and, to some extent, determine appropriate next steps compared to unaided internet search. The second study used mixed methods to compare how people use AI tools on their own skin concerns versus getting information from physicians. The research is careful to note that improving condition identification does not automatically translate to better decision-making about next steps — a known limitation of prior non-AI tools. Google's framing emphasizes human factors alongside technical capability, which is the right lens for consumer health AI. The practical implication is that AI-assisted dermatology information tools can provide measurable benefit to users who lack specialist vocabulary — someone searching for 'red dots on legs' might not know to search for 'palpable purpura,' and AI bridges that gap.