The Global Risks Alliance (GRA)
Topics at AI and Financial Model Risk: the pages in this section are listed below, each with its own scope and status.
The AI and Financial Model Risk topics are the eight subject dossiers of its field, among them model risk governance, credit and pricing models, and generative AI in finance. Each dossier sets out how the risk develops, what is measured and in which units, the public sources and their licences, the methods, who decides and the limits of the evidence.
GRA is a nonprofit business league for the financial services industry. It does not lend, underwrite, rate, invest, advise on or arrange transactions, and readiness work is not approval.
Topic pages in AI and Financial Model Risk share one structure: mechanisms, indicators, sources, methods, decisions and limits, written for model risk managers and validators and data scientists in finance.
Each term as this page uses it. Where a standard-setting body defines the term, the wording follows that body.
| Term | Definition |
|---|---|
| Model tiering | Classifying models by materiality and complexity to set how deeply each is validated and monitored. |
| Probability of default | The estimated likelihood that a borrower defaults within a set horizon, usually one year. |
| Hallucination | Plausible but false or unsupported content generated by a language model. |
| Feature attribution | A method that estimates how much each input contributed to a model's output. |
| Data drift | A change in the distribution of a model's inputs compared with the training data. |
The bodies and documents this page is read against. Status and editions change; each link goes to the issuing body.
Inventory, ownership and validation.
TopicFinance-specific automated decisions.
TopicUse cases and failure controls.
TopicInterpretation, bias and accountability.
TopicChanges over time and review thresholds.
TopicAuthority and approvals.
TopicVendor dependencies and due diligence.
TopicVerification and uncertainty.
The current service level is R0: information and enquiries. An enquiry receives a reference and names the next step; applications, payments, compute allocations and registrations open only when their team, terms and review capacity exist.