On May 27, 2026, the Tongyi Qianwen app under Alibaba announced an important upgrade to its "Photo Health Inquiry" feature. The new version breaks through the traditional OCR text recognition limitations and achieves medical image understanding and clinical reasoning capabilities.

d791a951-b0b2-4c86-9193-01148eaaeb48.png

Core Function Upgrade

  • Visual Selection: After users upload lab reports or photos of affected areas, AI can directly circle abnormal areas on the image (such as arrow indicators, rash details), just like a doctor would.
  • Integrated Reasoning: Instead of interpreting individual indicators in isolation, it now performs exclusion analysis based on age, gender, and multiple indicators. For example, it can identify that elevated ALP in adolescents is physiological rather than pathological.
  • Transparent Process: The system autonomously accesses the medical knowledge base and demonstrates the entire process of "recognition - annotation - reasoning," allowing users to understand the logic behind the conclusions.

Professional Clinical Validation Three senior physicians from Peking Union Medical College Hospital, Friendship Hospital, and Peking University People's Hospital tested the function. In multiple cases such as abdominal pain in a 13-year-old, possible fatty liver, urticaria with high white blood cell count, and chin rash, Qianwen's analytical approach (such as excluding biliary issues and differentiating between physiological and pathological elevations) was recognized by experts. They believe that it has diagnostic differentiation awareness and can effectively reduce user anxiety caused by misreading indicators. This upgrade marks a key step for AI health assistants moving from "mechanical value reading" to "auxiliary diagnosis," providing users with more reliable pre-check references.