Google is expanding its medical AI research beyond diagnosis and into long-term care management, according to a study published in Nature.
The company said its Articulate Medical Intelligence Explorer, or AMIE, has been adapted to help with the ongoing management of health conditions, not just one-time clinical conversations. That shift matters because many patients require repeated follow-ups, medication adjustments and updates based on changing guidelines after an initial diagnosis is made.
In Google’s description of the work, AMIE uses Gemini models with long-context capabilities to handle two parts of the task. One part is an empathetic dialogue system that can converse with patients in real time. The other is a reasoning agent designed to review extensive medical information, including drug formularies and clinical guidelines.
Google said the system can cross-reference hundreds of pages of clinical material while supporting the conversation layer. The goal is to help answer questions that arise over time as a condition develops, rather than only during an initial visit.
The research was tested in a blinded study using patient actors. Specialist physicians compared AMIE against 21 primary care doctors. According to Google, AMIE performed at a level comparable with clinicians in overall management reasoning. The company also said the system scored higher on plan precision and alignment with medical guidelines.
Google framed those results as evidence that AI may eventually help support routine medical care, with the aim of giving physicians more time to spend with patients.
The study does not mean AMIE is ready to replace doctors, but it does point to a broader role for AI in healthcare workflows. Google’s focus is now on how such tools might assist with ongoing treatment decisions, medication management and the interpretation of evolving clinical recommendations.
That direction reflects a larger challenge in medicine. Diagnosing a condition is only the beginning. Patients often need repeat assessments, symptom tracking and adjustments to care plans across many appointments. Systems that can remember prior context and weigh guideline changes could help clinicians manage that load.
Google said it is now exploring how AMIE could function in clinical settings. The company also said it has launched a nationwide study to evaluate AI in real-world virtual care.
Those next steps will likely be closely watched by both researchers and clinicians, especially as AI tools face scrutiny over reliability, safety and how they fit into existing care systems.
For now, the new research marks a clear attempt to move medical AI into a more sustained role. Instead of focusing only on identifying a condition, Google is positioning AMIE as a system that could help follow that condition through treatment and monitoring over time.