According to OpenAI, the integration currently provides read-only access, meaning the artificial intelligence can analyze records but cannot write or save new information back into the medical files.
This move aims to streamline clinical workflows by allowing medical staff to synthesize complex patient histories and clinical timelines without leaving their primary software.
To support these professional tasks, OpenAI also introduced a Healthcare Public Data plug-in that pulls information from external sources like PubMed and ClinicalTrials.gov to assist with trial eligibility and medication identification.
The company is now permitting healthcare organizations with a Business Associate Agreement—a legal contract required to handle protected health information—to use its specialized AI tools and connectors for compliant data processing.
Despite these deep technical integrations, OpenAI maintains that its software is not designed for diagnosis or treatment.
The company reported that while 99.1% of physician-reviewed AI responses were deemed safe in internal testing, it continues to face legal challenges over the accuracy of its health advice.
These concerns include recent lawsuits alleging that the AI provided dangerous recommendations regarding medical dosages, highlighting the ongoing risks as the technology is adopted across the healthcare industry.