The Clinical AI
Charter
Who is responsible when AI participates in care, and what rules does it follow?
Medicine has established ethical and professional standards for physicians, nurses, pharmacists, and other members of the care team. Clinical AI that performs many of the same functions should be governed by comparable principles.
.avif)
The Clinical AI Charter is an independent consensus framework, developed by physicians, ethicists, and health system leaders, that defines how Clinical AI should behave in the care of patients.
.avif)
UpDoc initiated the Charter and contributed its founding draft, but the document itself belongs to the field.
It was funded by [FUNDER], developed through a consensus process led by [METHODOLOGIST] at [INSTITUTION], and authored collectively by a panel on which UpDoc held one seat among many. We committed in advance to follow whatever the panel ratified, and we do.
.avif)
It began with a draft we invited the field to challenge.
UpDoc built the first agentic Clinical AI system that went through the FDA, and the commitments engineered into that system became a founding draft of principles. Rather than publish them ourselves, we brought them to an independent process and asked the field to test, revise, and decide what Clinical AI should be held to.
The process was led independently.
[METHODOLOGIST NAME, TITLE] of [INSTITUTION] designed and directed a formal Delphi consensus process, sponsored by [NAME OF SPONSOR]. A multi-institution panel of practicing physicians, health system leaders, researchers, and clinical ethicists deliberated, rated, and revised the principles across structured rounds. UpDoc participated as one of [N] panelists, contributing technical education on how Clinical AI systems work, with the same single voice as every other member.
The result is a ratified standard.
The consensus Charter was published in [MONTH YEAR]. UpDoc was its first signatory, and our platform upholds its principles.
Scale the pinnacle of medical practice, without compromise.
The promise of Clinical AI is consistency of quality: the ability to deliver the same standard of care that the best clinician in the world would deliver, to every patient, every time.
Our protocols are evidence-based, developed with clinical experts, and designed around patient outcomes.
Clinical AI acts within the scope of physician intent.
No Clinical AI action occurs without coordination with, and approval from, the patient’s physician. Whether interacting directly with the patient or operating in the background, Clinical AI must remain fully integrated with and supportive of clinical care.
Every action the system can take is defined, bounded, and authorized within the care plan the physician sets.
Safety before optimization.
Clinical AI does not optimize for efficiency, engagement, or cost metrics at the expense of clinical safety. When those objectives conflict, safety governs.
We enforce this principle in the core technical architecture of the system rather than leaving it to discretionary judgment.
When Clinical AI doesn’t know how to act, it escalates.
Clinical uncertainty is not a failure state. It is one of the most important signals the system can generate. Clinical AI must recognize the boundaries of its own confidence and route to a qualified clinician when those boundaries are reached.
Escalation is a designed feature of the system, and we treat confabulation as a safety defect rather than a tolerable error rate.
Patients are always aware when AI is participating in their care.
Clinical AI does not deliver care covertly. Patients are informed that an AI system is part of their care team, what it does, and what it cannot do. Transparency to the patient is a foundational requirement rather than a disclosure checkbox.
Patients are told, in plain language, when they are interacting with the system and how their physician remains involved.
Clinical AI must integrate into the clinician’s existing workflow.
Workflow integration is a clinical safety requirement rather than a product feature. An AI that clinicians route around will be ignored, and an ignored safety system is a failed safety system.
The system acts inside the care loop, working within the tools and processes care teams already use.
Every UpDoc action is auditable. There is always a why behind each decision.
Accountability cannot be diffuse. When Clinical AI acts, that action must be logged, traceable, and attributable, with the reasoning behind it preserved.
Every decision the system makes produces a reviewable record, a standard we hold ourselves to beyond what regulation requires.
Who wrote the Charter
The Charter was authored collectively by an independent, multi-institution consensus panel led by [METHODOLOGIST] of [INSTITUTION]. The panelists are practicing physicians, health system executives, researchers, and clinical ethicists, and each participated as an individual expert rather than as a representative of any company or institution.
How UpDoc treats patient data
The Charter sets the field's principles. This section is our own commitment, made in the same spirit. We treat patient data the way a responsible academic health system does: held in trust, used to advance care, and never treated as inventory. Here is what that means in plain English.
We will never sell identifiable patient data to anyone.
We will never share patient data with advertisers or advertising platforms.
We will never use patient data for purposes we have not disclosed.
We will never access more data than necessary to deliver safe care.
De-identified data makes the system better at caring for people like you. Learning from care in aggregate is how quality improves, and a system that cannot learn cannot get safer.
Clinical research happens with consent and oversight. When data is used for research, including clinical trials, it is either de-identified to HIPAA standards or used with the patient's explicit consent, under appropriate ethical and regulatory oversight. This is the same standard responsible health systems have followed for decades.
Health system partners receive aggregate quality reporting. The clinicians and institutions responsible for your care see how the system is performing, because they are accountable for it.
We publish our full data use policy in plain English, because we would rather explain all of this clearly than leave it to a terms of service document.
Built to be handed to your compliance team.
The Clinical AI Charter and its published methodology are available as standalone documents for health system legal, compliance, and clinical leadership teams.