Federal (Canada) · OSFI and Financial Consumer Agency of Canada (FCAC)
OSFI–FCAC AI risk report
OSFI-FCAC Risk Report – AI Uses and Risks at Federally Regulated Financial Institutions
The OSFI–FCAC report on AI risk, published September 24, 2024, sets out the risks Canada’s two federal bank supervisors expect banks to manage: bias in credit and pricing, keeping humans in the loop, being open with customers, and AI making scams such as deepfakes and voice spoofing cheaper to run. It is not formal guidance.
Figures as of November 1, 2026.Mitey, data editor · Evidence checked Status checked against the official sources listed below.
Published September 24, 2024. Not formal guidance, but it sets out the risks supervisors expect banks to manage.
Next milestone
No follow-up edition announced
The editor’s read
Why does it matter for AI in banking?
This report is the best single statement of what Canadian bank supervisors worry about with AI. Several of this Index’s criteria, such as telling customers when AI is used, follow its themes.
Read today, the report is a map of what came next: E-23’s extension to AI models, OSFI’s 2026 bulletins on generative and agentic AI, and FCAC’s consumer page. Its customer themes, transparency and humans in the loop, are the same ones this Index scores, and its warning about AI-enabled fraud is borne out by the proposed anti-fraud regulations and by the fraud and identity tools banks now disclose.

What are the key dates?
Report published
What does it mean for bank customers?
Regulators have told banks to watch AI for bias in credit and pricing decisions, keep humans in the loop and be open with customers about how AI and their data are used. It also warns that AI makes scams such as deepfakes and voice spoofing cheaper to run.
What does it mean for banks?
The report sets the supervisory baseline: about 50% of institutions used AI in 2023 with 70% expected by 2026, and risks span data governance, explainability, third-party concentration, cyber and AI-enabled fraud. Recommendations include treating AI as a cross-cutting risk, engaging boards and having an AI strategy even if not adopting AI.
Which institutions in the Index does it touch?
Who already tells customers about AI or automated decisions.
| Institution | Evidence | Index score |
|---|---|---|
| TD | Tells customers | 84 |
| RBC | Tells customers | 81 |
| CIBC | Tells customers | 80 |
| National Bank | Tells customers | 69 |
| BMO | Not found | 57 |
| Scotiabank | Tells customers | 55 |
| Desjardins | Tells customers | 54 |
| EQ Bank | Tells customers | 42 |
| Wealthsimple | Tells customers | 37 |
| ATB | Not found | 30 |
| Meridian | Not found | 24 |
| Neo | Tells customers | 24 |
| Tangerine | Tells customers | 21 |
| Coast Capital | Not found | 17 |
| Interac | Not found | 13 |
| Questrade | Tells customers | 13 |
| Simplii | Tells customers | 13 |
| KOHO | Not found | 11 |
| Canadian Tire Bank | Tells customers | 10 |
| Laurentian Bank | Not found | 8 |
| Manulife Bank | Not found | 5 |
| Beem | Not found | 3 |
| Vancity | Not found | 3 |
What does it ask for, and who already publishes it?
Each line paraphrases one expectation from the official sources. Where the Index measures something close, the last column counts the institutions in scope (who already tells customers about ai or automated decisions) that publish it, each linked to its source on their profile.
| What it asks | Closest Index criterion | Institutions that publish it |
|---|---|---|
| Watch for bias in credit and pricing decisionsA supervisory concern, not yet a rule. | Live customer use: Credit & lending | 5 of 23TD, RBC, CIBC, Neo, Vancity |
| Keep humans in the loop for material decisionsNamed by both supervisors. | Discloses board or committee oversight of AI | 5 of 23TD, RBC, CIBC, National Bank, Meridian |
| Be open with customers about AI and data useNamed by both supervisors. | Tells customers when AI or automated decisions are used | 13 of 23TD, RBC, CIBC, National Bank, Scotiabank, Desjardins, EQ Bank, Wealthsimple and 5 more |
| Defend customers against AI-enabled scamsThe report warns deepfakes make fraud cheaper. | Live customer use: Fraud & identity | 11 of 23TD, RBC, CIBC, National Bank, BMO, Scotiabank, EQ Bank, Wealthsimple and 3 more |
The checklist above, counted from the published Index data. Every number can be rechecked from the criteria CSV.
| Institutions in scope | 23Who already tells customers about AI or automated decisions |
|---|---|
| Checklist items | 44 matched to an Index criterion |
| Items published, across all institutions in scope | 34 of 92One point for each institution that publishes each measured item |
| Most items published by one institution | TD: 4 of 4 |
| Share of measured items published | 37%3 of 23 institutions publish every measured item |
What do the key terms mean?
- Automated decision
- A decision about you made by a system without a person reviewing it, such as an instant credit-limit decline.
- Right to an explanation
- A legal right to be told the main reasons and data behind a decision. Quebec has one; federal law does not yet.
What this means for you
- It is guidance or a report, not a binding rule.
- To see what your own bank has disclosed about AI and your rights today, use What AI touches my account?
- To complain about an AI-assisted decision, start with your bank’s complaint process; our guide on complaining about a bank AI decision has the steps.
Questions about OSFI–FCAC AI risk report
What is the status of OSFI–FCAC AI risk report?
Published September 24, 2024. Not formal guidance, but it sets out the risks supervisors expect banks to manage.
Did the report lead to new rules?
Indirectly. Binding change came through OSFI Guideline E-23, which covers AI models from May 1, 2027; the report itself remains guidance on risks supervisors expect banks to manage.
Is the report binding?
No. It describes risks and expectations; binding rules come through guidelines such as E-23.
Sources (official first)
Change log
Status verified against the official sources: Published September 24, 2024. Not formal guidance, but it sets out the risks supervisors expect banks to manage.
Related rules
- Bill C-36 (privacy, automated decisions) · proposed
- National AI Strategy and transparency consultation · proposed
- Bill C-34 (digital safety) · proposed
- Privacy commissioners’ generative AI principles · voluntary
Status as checked on November 1, 2026. Laws and guidance change; check the official source before relying on a date.
