What does SAMA Counter-Fraud Framework say about AI in banking?

Published Oct 11, 2022 · Last reviewed Nov 19, 2026

SAMA's Counter-Fraud Framework for the banking sector, sent to banks with Circular No. 44021528 on 11 October 2022, is the main binding SAMA text this review found that sets a requirement on AI. Where machine learning or artificial intelligence is used in counter-fraud systems, the system 'should not be black box' and should be capable of being audited, so that the bank can test what the algorithms are designed to do and whether they are correctly implemented. Banks with higher fraud risk are expected to add AI and machine learning to their detection systems, and every bank had to comply fully by 29 June 2023.

OFFICIAL TEXT: rulebook.sama.gov.sa ↗ · IN FORCE · SAUDI ARABIA (SAMA / SDAIA)

Checked 2026-11-19 · Saudi Arabia (SAMA / SDAIA), rulebook.sama.gov.sa
DocumentSAMA Counter-Fraud Framework — Counter-Fraud Framework (Saudi Central Bank, banking sector)
Issued bySaudi Central Bank (SAMA), the banking supervisor, with the Saudi Data and Artificial Intelligence Authority (SDAIA), the national reference for data and AI regulation
TypeFramework
StatusIn force
PublishedOct 11, 2022
EffectiveJun 29, 2023
Applies toBanks operating in Saudi Arabia (scope of application: banking sector). Binding: the covering circular requires banks to comply with its provisions, with full compliance due by 29 June 2023
Official sourcerulebook.sama.gov.sa ↗
Use casesFraud detection · Model risk management · Third-party & vendor AI · AI governance (general)

What are the key points of SAMA Counter-Fraud Framework?

  • Section 3.8 (Counter-Fraud Technology), control d.8: where machine learning or artificial intelligence is used, the system 'should not be black box' and should be capable of being audited, for example by being able to test 'what the algorithms are designed to do and whether they are correctly implemented'.
  • The same control list applies whether a fraud system is sourced from a vendor or developed in-house: the Counter-Fraud Department takes part in design and implementation with oversight from the CFGC, the rationale for scenarios and thresholds is documented, the bank can explain the fraud threats each scenario targets, and configuration changes follow SAMA's IT Governance Framework.
  • Section 5.2 (Fraud Detection Systems): where the fraud risk assessment finds higher risk or more fraud occurs, banks should add big data mining, predictive analytics and anomaly detection, and overlay 'Artificial Intelligence and Machine Learning algorithms (e.g., decision trees, random forests, neural networks)' to enhance decision-making, predict the likelihood of fraud and learn from historical patterns of fraudulent and legitimate behaviour.
  • Fraud detection systems should operate 24/7, draw on data across products, contact channels, external sources and intelligence monitoring, and be protected by data quality controls such as data governance, regular audit and regression testing.
  • Timeline in the covering circular: monthly gap-assessment reports from 30 November 2022, a roadmap to at least the third maturity level submitted by 30 November 2022 and approved by the board, full compliance by 29 June 2023, and an annual internal audit report on compliance from the end of Q4 2023.
  • The framework replaced the counter-fraud framework issued under Circular No. 41071315 dated 27/12/1441H, effective 29 June 2023.

What did SAMA Counter-Fraud Framework change for banks?

The 2022 framework replaced SAMA's earlier counter-fraud framework and treats AI in two ways: as a detection capability expected of banks with higher fraud risk, and as a control requiring that AI and machine learning fraud systems be auditable rather than a black box. It is the binding SAMA requirement on AI that applies to banks; SAMA has issued no AI-specific rule.

What does SAMA require when banks use AI for fraud detection?

Section 3.8 of the Counter-Fraud Framework says that where machine learning or artificial intelligence is used, the system should not be a black box and should be capable of being audited, for example by being able to test what the algorithms are designed to do and whether they are correctly implemented. The requirement applies whether the system is sourced from a vendor or developed in-house.

When did the SAMA Counter-Fraud Framework take effect?

SAMA sent it to banks with Circular No. 44021528 dated 11 October 2022 (15/3/1444H). Banks had to report monthly gap assessments from 30 November 2022 and comply fully by 29 June 2023, the date from which it replaced the framework issued under Circular No. 41071315.

DateDocumentStatus
Jul 14, 2026SDAIA National AI Risk Management Framework — National Artificial Intelligence Risk Management Framework (Executive Summary)Final
May 8, 2026SDAIA Deepfakes Guidelines — Deepfakes Guidelines: Mitigating Risks While Fostering InnovationFinal
Oct 15, 2025SAMA Governor statement on AI at the G20 FSI session — SAMA Governor Speech at the Financial Sector Issues (FSI) Session during the 4th G20 Finance Ministers and Central Bank Governors (FMCBGs) Meeting, Washington, D.C.Final
Jan 11, 2024SDAIA Generative AI Guidelines for Public — Generative Artificial Intelligence Guidelines for PublicFinal
Nov 19, 2026BNM AI Feedback Statement — Feedback Statement: Artificial Intelligence (AI) in the Malaysian Financial SectorFinal
Nov 19, 2026EBA Risk Assessment Report (Nov 2026) — Risk Assessment Report of the European Banking Authority, November 2026, with Box 1 on artificial intelligence related risks and opportunitiesFinal

Follow every document these regulators publish

when one of these regulators moves, the next morning's brief says so · six sourced stories · 7 am PT · free

plus every regulation, bank and agent page update, the morning after · leave any morning