Ontario AI Scribes Botch Drug Names: 60% Error Rate in 2026 Audit
A bombshell audit reveals that 60% of Ontario-approved AI scribes recorded the wrong prescribed drug during testing. The province’s flagship program to reduce paperwork is now under fire for prioritizing vendor presence over clinical accuracy.

Ontario AI Scribes Botch Drug Names: 60% Error Rate in 2026 Audit
summarize3-Point Summary
- 1A bombshell audit reveals that 60% of Ontario-approved AI scribes recorded the wrong prescribed drug during testing. The province’s flagship program to reduce paperwork is now under fire for prioritizing vendor presence over clinical accuracy.
- 2Ontario AI scribes, part of a flagship program to reduce physician paperwork, have been dealt a severe blow.
- 3A new report from the province's Auditor General, Shelley Spence, reveals that the AI note-takers routinely blow basic facts — including the names of prescribed drugs.
psychology_altWhy It Matters
- check_circleThis update has direct impact on the Etik, Güvenlik ve Regülasyon topic cluster.
- check_circleThis topic remains relevant for short-term AI monitoring.
- check_circleEstimated reading time is 4 minutes for a quick decision-ready brief.
Ontario AI scribes, part of a flagship program to reduce physician paperwork, have been dealt a severe blow. A new report from the province's Auditor General, Shelley Spence, reveals that the AI note-takers routinely blow basic facts — including the names of prescribed drugs. According to the audit, obtained by Orillia Matters, 60% of the AI scribe systems that made it onto the government's approved Vendor of Record (VOR) list recorded a different drug than the one actually prescribed during simulated patient encounters. The findings, released Tuesday, have ignited fierce debate about the safety and oversight of AI in clinical settings.
AI Scribe Errors in Ontario: Auditor General's Findings
The Auditor General's office tested 20 approved AI scribe vendors by having them transcribe two simulated conversations between healthcare workers and patients. The results were alarming. Beyond the drug-mixing errors, the report states that 17 out of 20 approved scribes “missed key details about the patients’ mental health issues in at least one of the two tests.” Even more troubling, nine of the 20 systems “fabricated information and made suggestions to patients’ treatment plans, such as referring the patient for therapy or ordering blood tests, even though these steps were not mentioned in the simulated recordings.”
AI Hallucinations: A Risk to Patient Safety
These AI hallucinations — where the software invents medical facts — represent a serious risk to patient safety. As noted by blog.jmir.org, clinicians are instructed to remain “in the loop” to review drafts for accuracy. However, the systemic nature of the errors uncovered by the audit suggests that human oversight may not be catching every dangerous mistake.
Supply Ontario Procurement Issues: Accuracy Weighted at Just 4%
Critics are pointing to the province's procurement scoring methodology as a root cause of the problem. The Auditor General's report reveals that Supply Ontario, the Crown agency that ran the bidding, weighted “accuracy” at a mere 4% of the total score. In stark contrast, “presence in Ontario” accounted for 30% of the evaluation criteria. This imbalance, experts argue, prioritized vendor market share over clinical safety.
The Impact on Physicians and Patients
“This is a fundamental failure of due diligence,” said a healthcare technology analyst who spoke on condition of anonymity. “You are buying a medical device that writes the patient record. If it can’t get the drug name right, the rest of the scoring is irrelevant.”
For physicians, the audit creates a legal and ethical minefield. The Ministry of Health's own Physician Fee-for-Service Post-Payment Audit Process makes clear that doctors are ultimately responsible for the accuracy of their billing and clinical records. If an AI scribe hallucinates a treatment plan or swaps a drug name, the physician — not the software vendor — faces potential audit clawbacks or malpractice liability.
Patients are also at risk. A drug mix-up in a patient's chart could lead to dangerous prescribing errors. The fabrication of mental health notes could stigmatize patients or lead to inappropriate referrals. The Auditor General's findings underscore that while AI scribes may reduce “pajama time” — the after-hours charting that burns out doctors — they must not introduce new, preventable hazards.
Medical Note-Taking Errors: What Happens Next?
The Ministry of Health and Supply Ontario have not yet publicly responded to the audit's specific findings. However, the report recommends an immediate review of the VOR list and a re-weighting of procurement criteria to prioritize clinical accuracy. The Ontario AI Scribe Program FAQs previously assured clinicians that the program “minimizes exposure to risk, with pre-negotiated contracts.” That assurance now rings hollow.
As Ontario pushes toward its goal of connecting every resident to a family doctor by 2029, the role of technology is critical. But as this audit makes clear, Ontario AI scribes must be held to the highest standard before they can be trusted with patient lives.

