Function has launched a secure health data connector that allows members to share their own lab results and clinician-reviewed notes with major generative AI platforms, including ChatGPT, Claude and Perplexity, the company said on Tuesday. The move is intended to make AI responses more personalised by using an individual’s biology as context for health questions.
What the connector does
The connector enables members to bring a range of personal clinical data into AI chats, including longitudinal health records, clinician-reviewed notes and more than 160 laboratory results. Function said this will help AI tools move beyond population-level generalities to answers that reflect an individual’s specific biomarkers and medical history.
"AI can empower us to understand and take action in so many aspects of our lives. Yet when it comes to health, AI has been missing the context that matters most: our biology," said Ziad Sultan, Chief Product and Technology Officer at Function.
Function said its service builds on its Private AI Chat offering, delivered through its Medical Intelligence Lab™, which members have used since last year to explore health questions. The company reported that more than 80% of Function members have at least one biomarker outside what it describes as an optimal range, citing examples such as elevated ApoB, low ferritin and abnormal thyroid markers as common findings that prompt urgent questions.
Intended uses and user scenarios
The company highlighted practical situations where the connector might be used: preparing for a cardiology appointment when there is a family history of heart attacks; interpreting concerning laboratory results before seeing a specialist; or investigating the causes of persistent fatigue. Function said integrating personal data with AI aims to make responses more specific and clinically relevant.
- Platforms supported: ChatGPT, Claude, Perplexity (as named by Function)
- Data types: clinician-reviewed notes, longitudinal history, 160+ lab results
- Company feature: builds on Private AI Chat and the Medical Intelligence Lab™
Function describes the connector as "secure" and frames the rollout as a step toward safer and more personalised AI-assisted health conversations. The announcement did not include technical specifications of security architecture, nor did it detail whether or how data sharing is logged, audited or reversible by members.
Context and implications
Millions of people in the United States and around the world are already turning to AI for health-related questions, and Function’s connector reflects a broader trend of integrating personal health data with generative AI. Proponents say that when AI can access an individual’s lab values and clinical notes, it can provide more pertinent explanations and clearer suggestions for what to raise with a clinician.
At the same time, the shift raises questions about data governance, clinical responsibility and the boundaries between informational support and medical advice. Function emphasises clinician-reviewed notes and member control over sharing, but its announcement did not specify the legal or regulatory framework that will govern use of the connector across different jurisdictions or how clinicians should treat AI-generated interpretations in clinical decision-making.
Health professionals and patients should regard AI outputs as informational and complementary, not definitive. Anyone using AI to interpret test results or symptoms should be encouraged to consult a qualified clinician or visit a clinic for assessment rather than relying solely on AI responses.
Function’s expansion follows its earlier Private AI Chat launch. The company characterised the broader rollout as a response to what it called an “immediate” need for personalised interpretation of biomarkers that fall outside optimal ranges, though it offered no independent clinical validation data in the announcement.
As health systems and regulators worldwide consider how to govern AI in medicine, tools that link sensitive clinical information to third-party AI models will draw scrutiny on matters such as consent, portability, de-identification and data security. Users and clinicians will need clarity on where responsibility lies when AI-generated suggestions influence care pathways.
For now, Function’s connector signals growing demand for AI that accounts for individual biology. Patients and clinicians in South Africa and beyond should watch closely for further detail on privacy safeguards, regulatory compliance and clinical validation before integrating such tools into routine care.
Note: Readers should consult a clinician or clinic for diagnosis and treatment — AI tools are not substitutes for medical assessment.
