Microsoft, Harvard Med partner for access to health info in Copilot

Microsoft might have some very close financial ties with Chat-GPT owner OpenAI, but the maker of Windows appears to be exploring ways to reduce its dependence on the wildly popular LLM. 

According to a report from The Wall Street Journal, Microsoft has struck a deal with Harvard Medical School to license consumer health content to Copilot, Microsoft’s own AI assistant. 

As a result, the latest version of Copilot, slated to launch later this month, will leverage content from Harvard Health Publishing to inform users about key clinical and wellness topics, avoiding the need to draw on OpenAI for answers to common health-related questions.  

Building trust in AI health information among everyday consumers

Chatbots like Chat-GPT are becoming major sources of health information for everyday users, with recent data from the Kaiser Family Foundation (KFF) estimating that just under 20% of adults use AI chatbots at least once a month for health information and advice. That number rises to 25% of adults under 30.  

However, trust in this information is pretty low, with only 29% saying they trust the tools to provide accurate info, and about one in five (23%) saying that AI is doing more to hurt those seeking accurate health information than to help them.  

While details of the Microsoft/Harvard Med partnership are scarce so far, The Wall Street Journal notes that Dominic King, vice president of health at Microsoft AI, said that one goal of the licensing deal is address this issue. The deal aims to infuse more detailed, accurate health information into Microsoft’s AI ecosystem and ensure that answers are closer to what users could expect from a licensed medical practitioner than the data that is currently available within the Copilot tool. 

“Making sure that people have access to credible, trustworthy health information that is tailored to their language and their literacy, and all kinds of things is essential,” he said. “Part of that is making sure that we’re sourcing that material from the right places.” 

Will healthcare accelerate the divergence of the Microsoft/OpenAI partnership?

The deal between Harvard Med and Microsoft also has some intriguing implications for AI industry observers – as well as executive decision-makers in healthcare organizations who are interested in adopting the best possible tools for their enterprise. 

At the moment, Copilot is largely known as a tool within Windows 11 and the Microsoft 365 suite able to assist with productivity tasks. Its foundations draw heavily on Chat-GPT models – no surprise, given that Microsoft has invested more than $13 billion into OpenAI and holds a 49% stake in its profits.  

The two entities still have a tight strategic partnership in place until 2030, but that hasn’t stopped both parties from looking into ways to pull away from each other and maximize profits in an AI environment that is still malleable enough for new and established entrants to be fighting hard for a piece of the pie. 

The problem for OpenAI is that it hasn’t actually turned a profit yet, even though it boasts around 800 million weekly active users, according to founder Sam Altman. The company has been sending out feelers in multiple directions to try to find a big score in profitability, most recently through budding enterprise partnerships.   

Some of those feelers are in the healthcare space.  But even though consumers are big (if uncertain) users of Chat-GPT, OpenAI doesn’t yet have a fully-fledged, stand-alone clinical product suite of its own.  

Even though even though OpenAI does offer HIPAA-compliant business associate agreements (BAAs) for healthcare organizations, clinicians may more likely to use Chat-GPT ad hoc for clinical decision support queries or to pay their dues to a third-party vendor that subsequently uses Chat-GPT as its AI engine.  

Meanwhile, Microsoft has firmer roots already established in the AI healthcare ecosystem. Not only do Bing and Copilot process 50 million health-related queries a day – and not only does Microsoft Azure cloud infrastructure support a huge amount of healthcare’s AI activities – but the company has also seen rapid adoption of both Dragon and DAX in the clinical setting: technologies that have been successfully nurtured after its acquisition of Nuance and a game-changing partnership with Epic Systems.   

Ambient listening technologies and documentation assistance tools have been some of the early big wins for AI in the direct clinical environment, and Microsoft has heavily touted its successes in improving access to patient care by enabling clinicians to stuff more appointments into their daily schedule.   

Boosting the quality of its consumer-focused clinical information in Copilot with Harvard Medical School’s data, potentially reducing the need for patients to take up their care team’s time at all for minor issues, could further its reputation as a both a consumer-facing resource and a clinical productivity powerhouse, giving it an even greater edge on both sides of the patient-provider relationship. 

If Microsoft pulls away in the healthcare arena, it could be the accelerant it needs to disentangle itself from its OpenAI frenemy and establish itself as a leader in both the consumer space and the enterprise market. 

Total success will likely require more than just licensing health information from Harvard, but it’s an interesting step into the world of consumer healthcare AI, where trust is proving to be a major sticking point.  

It will be fascinating to see how these two entities work to outmaneuver each other in healthcare and elsewhere even while being bound together in a complex web of financial and technological codependence. 

Jennifer Bresnick is a journalist and freelance content creator with a decade of experience in the health IT industry.  Her work has focused on leveraging innovative technology tools to create value, improve health equity, and achieve the promises of the learning health system.  She can be reached at [email protected].