Feature Request: Consent-Based Shared Knowledge and Longitudinal Patient Records Across ChatGPT Conversations
Dear OpenAI Product and Development Team,
I would like to propose a new ChatGPT capability that could make AI assistance significantly more useful for families, caregivers, researchers, and other people working together on a shared issue.
The problem
Today, different members of a family may independently discuss the same patient’s condition with ChatGPT. Over time, each conversation may accumulate valuable information about the patient’s medical history, symptoms, test results, medications, treatments, and responses to treatment.
However, this information may remain fragmented across separate accounts and conversations. When another family member starts a new conversation, ChatGPT may not have access to the relevant history, even when that history could help it provide more informed and consistent assistance.
As a result, users must repeatedly reconstruct the same history, and important context may be overlooked.
The proposed solution
I propose an optional, consent-based Shared Knowledge Space that allows authorized people to contribute to and access a common, continuously updated record associated with a particular person, case, or project.
For example, a family could create a shared patient record to which authorized members contribute relevant medical information and documents. When any authorized member consults ChatGPT about that patient, the assistant could use the shared record, subject to the user’s permissions and the applicable privacy safeguards.
The system could include:
-
A persistent, structured record: Organize relevant history, events, documents, test results, medications, and updates chronologically.
-
Controlled access: Let the record owner invite specific people and assign, review, or revoke their permissions.
-
Source tracking and reliability: Distinguish physician-confirmed diagnoses and documented test results from user-reported observations, hypotheses, and unverified information.
-
Cross-conversation continuity: Make relevant shared context available in new conversations without requiring users to repeatedly upload or explain the same information.
-
Conflict detection and updates: Identify inconsistent, outdated, or potentially conflicting information rather than silently treating every contribution as correct.
-
Privacy by design: Require explicit authorization, provide clear access logs and deletion controls, and prevent private conversations from being shared automatically.
-
Broader applications: Support other collaborative contexts, including caregiving, education, scientific research, technical troubleshooting, and long-term projects.
Why this matters
In a medical context, a coherent history can help users organize information, understand changes over time, prepare better questions for clinicians, and avoid repeatedly reconstructing the patient’s background.
The goal is not to replace physicians or allow unverified experiences to determine medical decisions. It is to help ChatGPT reason from a more complete and appropriately qualified context, while clearly communicating uncertainty and maintaining human oversight.
More broadly, this capability could transform ChatGPT from an assistant that primarily responds to individual conversations into an assistant that can support authorized collaboration around a shared, evolving body of knowledge.
An important distinction
This proposal is not about allowing ChatGPT to access other users’ private conversations without their permission. It is about providing a secure and transparent mechanism through which users can deliberately contribute selected information to a shared record and authorize others to use it.
I believe this could be a valuable direction for the future of ChatGPT, particularly for families managing complex, long-term situations.
Thank you for considering this proposal. I hope it can be evaluated as a potential product capability.
Sincerely,
A ChatGPT user