As of 2021, approximately 38.4 million Americans, or 11.6% of the population, were living with diabetes. This includes 29.7 million diagnosed cases and an estimated 8.7 million undiagnosed cases.

About 2 million individuals have type 1 diabetes, accounting for roughly 5% of all diagnosed cases. The remaining 90–95% of cases are type 2 diabetes.

Additionally, about 97.6 million U.S. adults have prediabetes, a condition where blood sugar levels are elevated but not yet high enough for a diabetes diagnosis. 

Patients with Type 2 diabetes now have access to a growing arsenal of digital health tools designed to help them manage their condition, including mobile apps, wearable devices, and remote monitoring systems. These tools allow patients to track blood glucose levels at home—typically using a glucometer—and share their data with their care team via apps, web portals, or during medical visits.

While these advancements have empowered many, they often assume a level of technological and mathematical proficiency that not all patients possess. Furthermore, they depend on the availability of healthcare professionals—a resource often stretched too thin.

A New Approach: AI Meets Familiar Technology

To address these challenges, researchers at Stanford Medicine have developed a voice-activated artificial intelligence (AI) application that operates on widely used smart speakers, such as Amazon Echo or Google Nest devices. This system provides real-time guidance for insulin dosing, eliminating the need for frequent doctor visits or phone calls.

“Access to care remains a significant barrier for many,” said Dr. Ashwin Nayak, a clinical assistant professor of medicine and co-author of the study. “This technology empowers patients to take control of their health.”

The groundbreaking system was tested in a study recently published in JAMA Network Open, co-authored by Dr. Nayak and Dr. Sharif Vakili. The results were impressive: patients using the smart speaker-based AI achieved optimal blood glucose levels more quickly and consistently than those using traditional methods. They also reported reduced emotional stress related to diabetes management.

Study Highlights: Simplified Insulin Management with AI

The small yet impactful study included 32 participants with Type 2 diabetes who were using insulin but struggling to maintain glycemic control. Half the participants were given a smart speaker preloaded with Stanford’s voice-based AI application, while the other half used standard methods of care.

For those in the AI group, the system personalized insulin recommendations based on each participant’s existing insulin protocol, target blood glucose range, and recent readings. By saying, “Check in with clinical trial,” participants engaged in a daily dialogue with the device, providing their fasting glucose levels and insulin usage. The AI would then offer tailored instructions, such as, “Increase your dose by X units,” or confirm that no changes were needed.

In contrast, participants in the standard care group used their smart speakers only for reminders to log their data and communicated with healthcare providers through phone calls or online portals when dosage adjustments were necessary.

Better Outcomes with Less Effort

The results demonstrated the transformative potential of this AI-powered approach:

• Glycemic Control: Over the eight-week trial, 81% of participants in the AI group achieved optimal blood glucose levels, compared to only 25% in the standard care group.
• Fewer Appointments: Patients in the AI group required significantly fewer doctor visits to adjust their insulin dosage, streamlining their care process.
• Improved Adherence: The AI-driven system encouraged more frequent and precise insulin dose adjustments, a critical factor in achieving glycemic control.
• Reduced Stress: Participants reported feeling less overwhelmed by the complexity of managing their diabetes.

“The integration of AI and voice interfaces has the potential to meet patients where they are, making diabetes management less burdensome,” said Dr. Kevin Schulman, co-director of Stanford’s Clinical Excellence Research Center and lead researcher on the study.

Expanding the Horizon: Broader Applications in Chronic Care

Although the trial focused on Type 2 diabetes, the researchers believe this technology could have far-reaching applications. The voice-activated system can be adapted for other chronic conditions, such as hypertension or heart failure, by monitoring patients’ responses to medications and adjusting treatment plans accordingly. The system is currently programmed in English but can be customized for other languages and healthcare systems.

Drs. Nayak and Vakili have also co-founded a startup, UpDoc, to commercialize the platform and bring this innovative technology to a wider audience.

A New Era of Patient-Centered Care

This study represents a pivotal step toward leveraging AI to address gaps in chronic disease management. By combining AI with an accessible and familiar technology, Stanford Medicine’s solution makes diabetes care more inclusive, efficient, and user-friendly—especially for older adults or those less comfortable with traditional digital tools.

As healthcare systems face increasing pressure to serve larger patient populations with limited resources, tools like this voice-activated AI platform may redefine how we approach chronic disease management, ensuring more patients receive the care they need when they need it.

Are you interested in how AI is changing healthcare? Subscribe to our newsletter, “PulsePoint,” for updates, insights, and trends on AI innovations in healthcare.