Skip to content

GRA · AI and Financial Model Risk

Spatial financeWhere capital meets hazard, exposure and transition

Third party models

Firms use vendor credit scores, fraud models, catastrophe models, market data analytics and foundation models they did not build and cannot fully inspect. Responsibility for model risk stays with the firm. This page sets out how third-party model risk arises, how firms validate what they cannot see, the evidence available, and who decides.

01नोंदणी करा

Third party models on the record.

02The question

Third party models: how the risk develops.

Vendor dependencies and due diligence. The dossier below sets out how the risk develops, what is measured, where the data come from, how it is modelled and who decides.

यंत्रणा

Limited transparency

Vendors protect their methods as trade secrets. Firms may get documentation and performance reports but not code, data or training details, which limits validation.

यंत्रणा

Shared models, shared errors

When many firms use the same vendor model, its errors affect them all at once and can move markets or credit supply together.

यंत्रणा

Updates change behaviour

Vendors update models and data on their own schedules. Without notice and change analysis, a firm’s validated model can change silently.

यंत्रणा

Validation by outcomes

Where internals are hidden, firms validate through benchmarking, outcome analysis and testing on their own data, which needs time and data.

03देखरेख करा

Model risk lifecycle.

The screen a team would use to follow this subject: its measures, states and decisions. Names are examples and values are illustrative.

Model risk lifecycle

Third party models · Model inventory · Committee 8

5 validations overdueGRA · AI and Financial Model Risk

Lifecycle

  1. Inventoryregistered
  2. Developdocumented
  3. Validateindependent
  4. Approvecommittee
  5. Implementcontrols
  6. Monitorthresholds
  7. Reviewannual
  8. Retiredecommission

Models

MD-709 PD retailTier 19 months1Overdue
MD-561 LLM complaints triageTier 1never0Not validated
MD-601 IFRS 9 stagingTier 13 months0मंजूर
MD-088 marketing propensityTier 314 months1मंजूर

Validation scope

संकल्पनात्मक भक्कमपणा
In progress
परिणाम विश्लेषण
Pending
Benchmarking
Pending
Use limits
human review required
Model risk lifecycle: the screen’s structure, terms and states for Third party models.Example names and illustrative values; a design, not a record or a live system.
04निर्देशक

What is measured, and how well.

Each indicator carries its unit, its cadence, the kind of source behind it and the evidence label a value would carry. No values are shown: a programme supplies them under its own authority.

IndicatorUnitCadenceSource typeलेबल
Third-party models in inventorycount by tierannualModel inventoryDECLARED
Vendor models with full documentationshareannualVendor management recordsDECLARED
Vendor model changes notifiedcountप्रति वर्षVendor noticesनोंदवलेले
Third-party models validated on own datashareannualValidation recordsDECLARED
Concentration of vendor model use across firmsshare of firms using top vendorsperiodicSupervisory surveysनोंदवलेले
Contracts with model change notice clausesshareperiodicContract reviewsDECLARED
05Exposure, loss and capital

What third party models means for exposure, loss and capital.

The exposure, loss, scenario and capital views a risk officer, supervisor or ministry of finance would ask for, with the frameworks they report against. Structures follow supervisory and actuarial practice. Values are illustrative: never prices, ratings or advice.

Transmission channels

How a shock in third party models reaches credit, market, underwriting, liquidity and operational risk, and feeds back.

Channel structure after the Basel Committee’s work on climate-related risk drivers and transmission channels; links shown are typical paths, not estimated exposures.
NGFS climate scenarios · Network for Greening the Financial SystemIFRS S2 Climate-related Disclosures · ISSBTCFD recommendations · Financial Stability Board TCFDPrinciples for the effective management of climate-related financial risks · Basel Committee on Banking Supervision
06स्रोत

Where the evidence comes from.

Named public datasets and institutional records, each with its licence class. Sources by agreement are used only under their terms and never republished.

स्रोतWhat it givesLicence class
SR 11-7 guidance on vendor and third-party modelsExpectations for using external modelsसार्वजनिक क्षेत्र
Interagency guidance on third-party relationships (United States banking agencies)Third-party risk management expectationsसार्वजनिक क्षेत्र
FSB report on AI financial stability implicationsThird-party and concentration risks from AIFree, FSB terms
EU AI Act obligations for providers and deployersDivision of duties between AI providers and usersFree, EU reuse terms
Vendor model documentation and performance reportsInformation vendors provide to usersProprietary, terms per vendor
07Decisions

Who decides, and what the evidence serves.

AI and Financial Model Risk prepares evidence. The decisions below are taken by the authorities named, under their own rules, and never by the platform.

Who decidesThe questionEvidence this page serves
Model risk functionsHow to validate and approve third-party modelsDocumentation, benchmarks and outcome testing
Procurement and legalWhich transparency and change terms to requireContract practice and regulatory expectations
SupervisorsWhether vendor model use is governed adequately and whether concentration mattersSurveys and review findings
VendorsWhat documentation and notice to provideCustomer and regulatory requirements

Outputs this topic produces

  • Guide to validating models firms cannot inspect
  • Brief on concentration in vendor and foundation models
  • Method note on change analysis for vendor updates
  • Questions for procurement on model transparency terms
08Core questions

What experts ask.

The questions concern responsibility for models one did not build.

  • Does the institution understand how a vendor model works well enough to validate its use?
  • What documentation, testing evidence and access does the vendor provide?
  • How are vendor model updates communicated, tested and approved?
  • How many institutions rely on the same vendor, and what happens if it fails or changes?
  • Can the institution exit or replace the model if needed?
09Mechanisms

Mechanisms and drivers.

Vendors often protect their models as intellectual property, limiting the transparency available to users. Institutions must still validate the model’s use for their own purposes, typically through outcome testing, benchmarking and contractual rights to documentation. Updates by the vendor can change model behaviour without the institution’s knowledge.

Concentration is a sector-wide concern. When many institutions use the same vendor model or AI provider, a flaw, outage or change affects them together, creating correlated errors and operational dependencies that no single institution can manage alone.

10Methods and data

How it is studied, and where the methods stop.

Method or dataWhat it showsमर्यादा
Vendor due diligenceVendor capability, controls and transparencyLimited by vendor disclosure
Outcome and benchmark testingPerformance on the institution’s own dataDoes not reveal internal logic
Contract and exit analysisRights to information, notification and exitRights may be hard to exercise
Concentration mappingShared dependencies across the sectorRequires aggregated information
12प्रश्न

प्रश्न आणि उत्तरे.

Who is responsible for a vendor model’s risk?
The firm using it. Supervisory guidance expects firms to validate and govern third-party models as they would their own, within the information available.
How can a firm validate a model it cannot see?
Through documentation review, benchmarking against alternatives, outcome analysis, sensitivity testing on its own data and contractual rights to information.
Why does vendor concentration matter?
Because a flaw or change in a widely used model affects many firms at once, which can correlate decisions across the market.
Does buying a model transfer its risk to the vendor?
No. The institution using the model remains responsible for its use and outcomes.
Why is vendor concentration a stability concern?
Because if many firms rely on the same model or provider, a single failure or change can affect them all at once.
What should an institution be able to obtain from a model vendor?
Typically enough documentation to understand the model’s purpose, design, data and limitations, evidence of the vendor’s own testing, notice of material changes, and rights that allow the institution to validate the model’s use and to exit. What is obtainable in practice varies.
Does the platform rate vendors?
No. GRA does not rate, endorse or recommend vendors.
अहवालयोगदान द्या

या पानावरून पुरावा योगदान द्या.

Reports is the Consortium’s quarterly publication. What you know about this subject can become evidence, a case or a correction in the next edition, credited to you and published under CC BY 4.0 where it is your own work.

Global Risks AllianceAI and Financial Model RiskMethods and TechnicalExponential Technology

वैयक्तिक डेटा, संवेदनशील पायाभूत सुविधांचा तपशील किंवा सामायिक करण्याचा तुम्हाला अधिकार नसलेले काहीही कधीही पाठवू नका. योगदानाचे वापरापूर्वी पुनरावलोकन केले जाते; ते पाठवल्याने कोणतेही लेखकत्व किंवा संपादकीय अधिकार मिळत नाहीत, आणि प्रायोजकत्वाने कधीही निष्कर्ष विकत घेता येत नाही.

AI आणि वित्तीय प्रारूप जोखीम

Bring an exposure or capital question to the Alliance.

Information and enquiries are open today. 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.

कायदेशीर

या साइटवरील नकाशे, डेटा आणि आकृत्या संदर्भ आणि सर्वसाधारण माहितीसाठी दाखवल्या आहेत. त्या सल्ला, अंदाज किंवा इशारे नाहीत; माहितीच्या मर्यादा अस्वीकरणे आणि डेटा सूचनांमध्ये नमूद केल्या आहेत.