Google Unveils Medical AI That Outperforms Doctors in Diagnosis
Google has developed medical artificial intelligence AMIE, which in tests showed higher diagnostic accuracy than experienced doctors. However, the technology is not yet ready for real clinical practice - more research is needed.

Google has unveiled a new medical AI system, Articulate Medical Intelligence Explorer (AMIE), which the company claims can outperform experienced doctors in the accuracy of diagnostics based on images. The development is based on the Gemini 2.0 Flash language model, which can process both text and visual data, allowing the AI to analyze photographs, document scans, and lab results. The system is trained to diagnose skin diseases, decipher ECGs, and interpret tests, potentially becoming a tool for telemedicine.
To test the effectiveness of AMIE, Google ran simulations involving 25 actors portraying patients with different symptoms. In 105 scenarios, AI and doctors made diagnoses based on the data provided, and independent experts assessed their accuracy. The results showed that AMIE performed better on average, especially with non-obvious cases. However, the developers emphasize that the tests were artificial, and larger-scale studies in clinical settings are needed before implementation in real practice. Experts emphasize that such systems should be used as auxiliary tools, and not replace doctors.
What's Your Reaction?






