Haryana's 112 Emergency Service — AI Auto-Dispatch Pilot Cuts Response Time Toward Under 10 Minutes
India
India's Haryana 112 helpline handled 2.75 crore (27.5M) calls over four years with 92.6% caller satisfaction and response times cut …
Canada · Toronto · See the Canada profile · See the Toronto profile
Top 62% 53/100 · Ask Evidence Copilot about this practice
Ontario approved 20 AI scribe vendors for physicians (5,000+ participating). The Auditor General's May 2026 testing found 9 fabricated content, 12 swapped drugs, and 17 missed mental-health details, while procurement weighted Ontario presence at 30% vs 4% for accuracy.
Ontario's AI Scribe programme, run through Supply Ontario with the Ministry of Health and OntarioMD, approved 20 vendor systems that transcribe patient visits and draft clinical notes. The Ministry of Health says more than 5,000 Ontario physicians participate, and OntarioMD attributes paperwork-time reductions of 70-90% to such tools (a programme claim from an earlier report).
In a special report on AI use in government tabled on 12 May 2026, the Auditor General of Ontario reviewed the vendor evaluation, which used simulated doctor-patient recordings compared against source material by medical professionals. Of the 20 systems, 9 fabricated information (for example invented referrals, test orders or findings such as 'no masses'), 12 recorded a different drug than the one prescribed, and 17 missed key mental-health details in at least one test (6 in both tests).
The audit also found that the procurement scoring gave 30% weight to a domestic Ontario presence but 4% to note accuracy, 2% to bias controls and 2% to threat, risk and privacy assessments; vendor test results were not validated, live demonstrations were limited, and none of the approved products required clinician sign-off attestation. The report made ten recommendations (five to the Ministry, five to Supply Ontario), including higher weights and minimum pass marks for accuracy, bias and security, mandatory live demonstrations, annual SOC 2 Type 2 reports and attestation controls.
The Ministry agreed to its recommendations and Supply Ontario agreed or partly agreed to theirs. The Minister said hallucinations arose in testing, not in operational use; the Ministry knows of no patient harm. Evidence limits: findings come from simulated recordings, not clinical outcomes, and no real-world error rates are published. The case is valuable as a transparent audit of how procurement weighting can let accuracy defects through.
Read the full analysis: https://www.hrreporter.com/focus-areas/automation-ai/ai-systems-for-ontario-doctors-hallucinate-auditor-general/394430
Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.
Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.
Where this practice's information was retrieved from, and when.
India
India's Haryana 112 helpline handled 2.75 crore (27.5M) calls over four years with 92.6% caller satisfaction and response times cut …
Guyana
Guyana's Ministry of Health and Mount Sinai Health System are rolling out AI-assisted radiology and pathology reads, a reopened NYC-linked …
Canada
Canada's tax agency has spent over $18M on its FAQ chatbot Charlie since 2018-19. In October 2025 the Auditor General …
Portugal
Portugal committed EUR 11.5M (2023-2027) to a ChatGPT-based 'Voicebot 112' meant to triage overloaded emergency calls, but contractor Sword Health …
Open full copilot Grounded in cited practices — always check the sources.