Ask a Scientist: How can AI help doctors spend more time on collaborative care?

Imagine showing up for an urgent care visit, and there’s no paperwork to fill out. Your doctor already has the details, because you chatted with an AI agent after booking your appointment. That agent went further than the one-size-fits-all questions you’d find on a standard intake form, asking you targeted follow-ups based on the personalized details you provided. Those specifics mean your doctor can spend less time information-gathering and more time on collaborative care: addressing your concerns, diagnosing your symptoms, and figuring out what to do next.

Google researchers are building that system. It’s called AMIE (Articulate Medical Intelligence Explorer). And recent studies have shown the difference it can make for patients and doctors alike.

We sat down with research scientists Anil Palepu and Mike Schaekermann to learn more.

What do you do at Google?

Anil: We’re both research scientists on the AMIE team. I do a lot on the technical side — evaluating the AI agent, preparing it for studies, and expanding it so it can have real-time conversations over audio and video, not just text.

Mike: My work mostly involves planning clinical studies, building partnerships with healthcare providers, and submitting our research to journals.

What is AMIE, exactly?

Mike: AMIE is Google’s flagship research project exploring the future of AI in healthcare. We envision what’s possible — and see if we can make that a reality.

In this case, we wanted to train an AI agent to have clinical conversations. At the outset of our research, we envisioned something you could talk to any time you wanted, as long as you wanted, about your health concerns. And it would act like a bridge between you and your doctor, carefully synthesizing and sharing the information you’ve given it. That way, at your next appointment, your doctor could maximize their time with you.

How can AMIE potentially change the average doctor’s appointment?

Anil: AMIE is more targeted and more thorough than a standard doctor’s office intake form.

Let’s say I come in with back pain. While an intake form might ask me to describe the problem and check a few boxes, AMIE can ask follow-ups. What kind of pain? When did it start? Where is it in my back, exactly? That gives the doctor a better understanding of why I’m there, giving them more time to dig into potential causes and work with me to develop a treatment plan.

Mike: Doctors have told us AMIE helps them shift from gathering information to collaborating more with patients. When doctors don’t have to spend so much of their time verifying baseline facts, they can focus on discussing treatment options and making decisions with the patient.

Tell me more about your research. How did it work?

Mike: Study participants had the option to chat with AMIE before their next doctor’s appointment. AMIE took their clinical history and shared a concise summary with their physician. We had a trained physician oversee every AMIE-patient interaction just in case something unsafe happened, and they never had to interrupt the conversation or intervene.

What’s next for AMIE?

Anil: We’re working on a follow-up study that builds on what we’ve done so far. We want to see how well AMIE works on a much broader scale, across all of the US, with even more patients and involving controlled comparisons.

Mike: We also want to expand AMIE so the patients in our studies can interact with it in more ways. AMIE started as a text chat. Then, patients could share pictures in that text chat. And now we’re exploring what’s possible with video, so we can test how it works when patients talk to it out loud and show it what they’re experiencing.

What do you hope people take away from your work?

Mike: We want to enable everyone, everywhere to live their healthiest lives. And we believe systems like AMIE can give physicians back crucial time with their patients. This was a chance for us to show what AMIE can do in a rigorous way and hopefully build trust with the medical community.

Anil: So far, the evidence has demonstrated benefits for patients and doctors — and potentially, for the healthcare system as a whole.