Microsoft’s Copilot Health can use AI to turn your fitness data and medical records ‘into a coherent story’

Microsoft has unveiled Copilot Health, an AI-powered tool it claims can help make sense of your medical records, health history and fitness data from wearables, should you grant it access to that information. The company said it will be in a “separate, secure space” in the Copilot app and that the idea is to help provide you with more context and insights so you can ask your doctor the right questions when you see them.

Copilot Health is designed to help you better understand your medical information as a whole, Microsoft says. It is not “intended to diagnose, treat or prevent diseases or other conditions and is not a substitute for professional medical advice,” the company pointed out in a blog post.

The tool can pull in activity, fitness and sleep data from more than 50 devices, including Apple Watch, Oura and Fitbit. Through HealthEx, it can access health records that include visit summaries, medication details and test results from more than 50,000 hospitals and provider organizations in the US. It can tap into lab test results from Function, should you allow it to do so.

Copilot Health can take all those details and apply “intelligence to turn them into a coherent story,” such as helping you pinpoint the reasons why you don’t sleep too well, the company suggested. It can access real-time provider directories in the US to help users find clinicians based on factors like location, specialty, spoken languages and insurance coverage.

Microsoft says that, across AI-powered consumer products like Copilot and Bing, users ask more than 50 million health-related questions every day. “We’ve improved the quality and reliability of answers by elevating information from credible health organizations across 50 countries, as verified by our clinical team using principles independently established by the National Academy of Medicine,” the blog post states. “Responses include clear citations with easy links to source material, alongside expert‑written answer cards from Harvard Health.”

As far as privacy is concerned, Microsoft says Copilot Health data and conversations are siloed from the broader Copilot app and there are extra access and safety controls in place, including “encryption at rest and in transit.” You can delete your information and cut off the app’s access to health records and wearable data at any time. Microsoft also notes that it won’t use your Copilot Health information to train its models.

The company explained that Copilot Health was informed by its responsible AI principles. Microsoft built the tool in collaboration with its own clinical team and with the expertise and feedback of more than 230 physicians from dozens of countries. “Copilot Health has achieved ISO/IEC 42001 certification, the world’s first standard for AI management systems, meaning an independent third party has verified how we build, govern and continuously improve the AI behind this service,” it noted.

Microsoft has opened up a waitlist for those interested in trying Copilot Health. The tool will initially be available in English in the US for those aged 18 and over. The company is working on adding support for more language and voice options and it will announce availability for those and other territories down the line.

While users will be able to try Copilot Health for free at first, Microsoft plans to charge for access via a subscription, according to The New York Times. The company has not yet disclosed pricing details.

The Copilot Health announcement comes just a couple of days after Amazon expanded its Health AI tool beyond One Medical. It’s now available on the Amazon website and app. Prime members in the US have the option to chat about certain conditions with a One Medical provider via direct message at no extra cost. Earlier this year, OpenAI announced that it was testing ChatGPT Health. Anthropic has healthcare tools as well.

Given how tough it is for many folks to access affordable healthcare and the fact their data and health records are often spread across a number of providers, some might believe there are benefits of using such tools from AI companies.

However, there’s a big difference between tracking your sleep or calling your doctor after an Apple Watch detects signs of atrial fibrillation and entrusting all of your medical information to a chatbot. There are also issues like AI hallucinations and chatbots providing users with straight-up bad advice, as well as the possibility that an LLM-based tool might downplay or exaggerate potential risks.