Medical Information AI Audit Checklist
Assess safeguards and oversight without entering patient or other sensitive information.
Overall, how would you rate the system’s safeguards for handling medical information?
*
1
2
3
4
5
How clearly are the system’s purpose and permitted uses documented?
*
Not clearly
1
2
3
4
Very clearly
5
1 is Not clearly, 5 is Very clearly
How consistently does the system minimize the medical information it processes?
*
Not consistently
1
2
3
4
Very consistently
5
1 is Not consistently, 5 is Very consistently
Which safeguards are documented? Select all that apply.
*
Access is limited by role
Data retention and deletion practices are defined
Data use and sharing are documented
Security testing or monitoring is documented
Incident response procedures are documented
None of these are documented
Other
How well are AI-generated outputs reviewed by qualified human reviewers before consequential use?
*
Not at all
1
2
3
4
Very well
5
1 is Not at all, 5 is Very well
How are system limitations and potential errors communicated to users?
*
Please Select
Clearly and consistently
Sometimes, but not consistently
Rarely or not at all
Not sure
How regularly are performance and safety risks evaluated after deployment?
*
Continuously or on a defined schedule
Occasionally, without a defined schedule
Only after an issue is reported
Not evaluated
Not sure
Rate each area based on the documentation or evidence reviewed.
*
Rows
Needs attention
Partially addressed
Well addressed
Not assessed
Data governance and accountability
1
2
3
4
Privacy and data minimization
5
6
7
8
Access and security safeguards
9
10
11
12
Output reliability and human oversight
13
14
15
16
Monitoring and incident response
17
18
19
20
What is the highest-priority improvement identified? Do not include patient details or sensitive information.
What additional non-sensitive context would help explain this assessment?
Submit Assessment
Should be Empty: