Microsoft Corp. is preparing to launch a major update to its artificial intelligence (AI) assistant Copilot that will incorporate medical information from Harvard Medical School, marking the company’s latest push to establish itself as a leader in healthcare AI.
The updated version of Copilot, expected as soon as this month, will draw on content from Harvard Health Publishing to answer user questions about medical topics, according to a Wall Street Journal report citing sources familiar with the matter. Microsoft will pay Harvard a licensing fee for access to the material.
While declining to discuss the Harvard partnership specifically, Dominic King, vice president of health at Microsoft AI, told the Journal that the company wants Copilot to deliver responses comparable to what users might receive from a medical professional.
“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,” King said. “Part of that is making sure that we’re sourcing that material from the right places.”
The initiative comes as AI chatbots face ongoing scrutiny over their reliability in providing medical advice. A 2024 Stanford University study found ChatGPT gave inappropriate answers to roughly 20% of 382 medical questions tested. King said the goal is to help users make informed decisions about managing complex conditions like diabetes. The company is also developing a feature that would help users locate healthcare providers based on their location, medical needs, and insurance coverage.
The Harvard Health Publishing data includes mental health content, though Microsoft has not disclosed how the updated Copilot will handle such sensitive queries. The handling of mental health issues by AI chatbots has drawn scrutiny from lawmakers and health experts, with ChatGPT allegedly playing a role in crises that resulted in hospitalization or death, the Journal previously reported.
The healthcare push comes as Microsoft seeks to reduce its dependence on OpenAI, despite the two companies’ close partnership. Mustafa Suleyman, CEO of Microsoft AI, has made healthcare a priority while expanding an internal AI lab that competes with OpenAI.
Microsoft formed a dedicated consumer AI and research unit in 2024 and is training its own models with the goal of eventually replacing OpenAI workloads, though the company acknowledges this could take years. Last week, Microsoft CEO Satya Nadella announced he would delegate some responsibilities to focus on the company’s most significant AI initiatives.
In August, Microsoft began publicly testing a self-built AI model for Copilot. The company has hired researchers and engineers primarily from Google’s DeepMind AI lab to advance its proprietary models. Microsoft is already using non-OpenAI models in some software products and has deployed models from OpenAI competitor Anthropic to power AI tools in its Microsoft 365 suite.
Despite the push for independence, Microsoft maintains OpenAI “will continue to be our partner on frontier models” and that its philosophy is to use the best available models. Microsoft and OpenAI signed a tentative agreement last month to redefine their relationship, paving the way for OpenAI to restructure as a for-profit company.
