source: google research: research into how ai can help users understand skin conditions
level: research
a large survey with over 2,300 people tested how an ai tool affects understanding of skin conditions. participants saw case images and history, then tried to name the condition and pick a next step. those using the ai tool were more willing to guess a name and got it right nearly three times more often than those using standard web search. a version with perfect ai predictions did even better, but accuracy was still far from perfect.
the ai tool showed a list of possible conditions with textbook images and details. this helped users feel more confident and satisfied with their search. however, picking the right next step, like whether to use a home remedy or see a doctor urgently, did not improve much. the standard ai did not show a significant gain, and users were slightly more likely to suggest a less urgent step than a dermatologist would.
a separate real-world study with 110 people from a diverse community used a similar app in four languages. participants could name their condition more often after using the app, and clinicians found the app's predictions matched their own views most of the time. the app also helped as a shared reference during doctor visits. both studies point to the need for tools that give more specific guidance and better visual examples across skin tones and conditions.
why it matters: ai can make health information more accessible, but designers must ensure it helps people take safe and correct actions, not just identify problems.
source: google research: research into how ai can help users understand skin conditions