Microsoft Copilot Health: Revolutionizing Medical Data
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Microsoft Unveils Copilot Health for Simplified Medical Data Management
Last Thursday, Microsoft unveiled Copilot Health, a new feature integrated into Copilot, designed to provide a secure space where users can ask questions about their lab results and medical records. This tool also allows for the analysis of data from wearable devices and other health-related discussions. The rollout of this feature will be gradual, and interested users can sign up for a waitlist to gain access.
Microsoft emphasizes that Copilot Health is not intended to replace a doctor and does not provide medical diagnosis or treatment. The goal is to help users better understand their health data. With HealthEx, users can import medical records from over 50,000 hospitals and health organizations in the United States. Additionally, lab test results can be imported via Function. Copilot Health is compatible with over 50 wearable devices, including those from Apple, Oura, and Fitbit. The main interface of Copilot Health can display data such as step counts, as well as reminders for upcoming medical appointments, depending on the information users choose to share.
Easier Access to Healthcare Professionals
Copilot Health also offers the ability to find healthcare professionals. It is connected to real-time directories of U.S. providers, allowing users to search for specialists based on criteria such as location, spoken languages, and accepted insurance plans.
In its announcement, Microsoft highlighted that it has improved the quality and reliability of the responses provided by Copilot Health by relying on information from credible health organizations in 50 countries. The responses include citations with links to reliable sources and "answer cards" written by experts from Harvard Health.
Data Security and Privacy
Microsoft assures that discussions in Copilot Health are isolated from the general Copilot and benefit from enhanced access controls, privacy, and security. The company guarantees that data from these discussions is not used to train its AI models. Users also have the option to delete their health data or disconnect data sources at any time, for example, by disabling access to data from wearable devices.
OpenAI launched a similar feature in January, called ChatGPT Health, which also provides a secure environment for medical discussions. However, Microsoft has not yet offered a version of Copilot Health that complies with HIPAA, unlike ChatGPT for health and Amazon's health AI, which has recently expanded its access. Anthropic's Claude for health is also "HIPAA-ready."
Compliance and Certifications
During a briefing, Dr. Dominic King, Vice President of Health at Microsoft AI, stated that HIPAA compliance is not necessary for a direct consumer experience using their own data. The Health Insurance Portability and Accountability Act imposes security requirements to protect electronic health data and prohibits certain uses and disclosures. Violations of HIPAA can result in fines and prison sentences. Although Microsoft is not legally required to comply with HIPAA, King emphasized the importance of adhering to the best standards in place and promised updates regarding Microsoft's position on HIPAA.
King also mentioned that Copilot Health has an ISO 42001 certification, an international standard for AI systems that promotes responsible AI use, as well as traceability, transparency, and reliability. Microsoft 365 Copilot and Microsoft 365 Copilot Chat also hold this certification.
Precautions to Take
Despite this certification and the intention to voluntarily comply with HIPAA in the future, users should remain cautious about sharing their medical data with AI. The data privacy policies of AI companies can change at any time. Additionally, AI has sometimes provided inaccurate or dangerous medical advice, particularly regarding mental health.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.