AI BriefingOntario Auditor Points Out Frequent Factual Errors in AI Note-Writing Tools for Doctors
Key point
An Ontario audit found that AI scribe systems for doctors showed serious hallucinations, including errors in medication information and missing patient information.
Details
An investigation by the Office of the Auditor General of Ontario, Canada found that a significant number of the 20 approved AI Scribe vendor systems were making serious errors.
Key error cases are as follows:
- 60% of systems confused prescribed medication information.
- 9/20 systems showed hallucination, suggesting treatment plans not present in the recording or fabricating information.
- 17/20 systems omitted key details related to patients' mental health.
The report pointed to inadequate evaluation criteria as one cause of these problems. In the vendor selection process, the accuracy of medical records accounted for only 4% of the total score, while having a business presence in Ontario accounted for 30%, meaning other factors were overly reflected compared to technical reliability.
As a result, warnings emerged that such an unbalanced evaluation method could lead to the selection of vendors with a risk of producing inaccurate or biased medical records.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Check Point Finds Jev Decision Model Vulnerable to Prompt Injection for ~$0.50Paper Instruments Releases Paper Office Python Suite to Boost AI Agent Document Editing Accuracy