The CSSF published 'Artificial Intelligence: Opportunities, risks and recommendations for the financial sector' on 21 December 2018, a research study that explains AI and machine learning, the model development life cycle and financial-sector use cases, and sets out key risks with recommendations. The paper says it 'has no binding value vis-à-vis the supervised institutions' but provides the foundations for dialogue with the sector. Its recommendations include that 'Accountability cannot be delegated to a machine', with ultimate responsibility for AI/ML solutions resting with senior management, together with human oversight, bias testing, explainability and audit trails.
OFFICIAL TEXT: cssf.lu ↗ · FINAL · LUXEMBOURG (CSSF / BCL)
| Document | CSSF White Paper on Artificial Intelligence (2018) — Artificial Intelligence: Opportunities, risks and recommendations for the financial sector (CSSF white paper) |
| Issued by | Commission de Surveillance du Secteur Financier (CSSF), Luxembourg's financial sector supervisor, with the Banque centrale du Luxembourg (BCL) on joint surveys of AI use |
| Type | Report |
| Status | Final |
| Published | Dec 21, 2018 |
| Applies to | Entities supervised by the CSSF, including credit institutions, investment firms, payment and e-money institutions and investment fund managers. A research study published as a white paper; the CSSF says it 'has no binding value vis-à-vis the supervised institutions' |
| Official source | cssf.lu ↗ |
| Use cases | AI governance (general) · Model risk management · Credit scoring & underwriting · AML / KYC · Fraud detection · Customer-facing chatbots · Cybersecurity · Third-party & vendor AI |
What are the key points of CSSF White Paper on Artificial Intelligence (2018)?
- Purpose: given 'the increasing adoption of AI in the financial sector and the relative lack of practical guidance from a risk perspective', the CSSF shared its study 'for the benefit of the financial sector'.
- Use cases covered: robotic and intelligent process automation, chatbots, robo-advisors, fraud detection and money-laundering investigations, terrorism financing, and credit scoring (automated loan decisions).
- Governance: depending on the criticality of the task being automated, 'implement a human oversight/dual control'; keep enough internal AI skills to understand, develop or supervise the solution, and make sure auditors can audit it.
- Bias and discrimination: implement an AI code of conduct, seek diversity in training and validation data, remove data bias during data preparation, create validation datasets to test against discrimination and monitor model performance continuously.
- Explainability and auditability: document data preparation and the model blueprint, choose more interpretable algorithms depending on the criticality of the system, and keep detailed audit logs from raw data to the AI/ML outcome.
- Security and robustness: the paper covers data poisoning, adversarial attacks, model stealing and video forgery, and recommends security by design and independent security reviews according to the criticality and exposure of the system, noting that 'attackers are also using AI to improve their attacks'.
- External providers: evaluate the maintenance needs of an AI/ML product, apply IT outsourcing rules (it cites Circular CSSF 12/552), and monitor systemic effects where the same model is used by many institutions.
- Accountability: 'Accountability cannot be delegated to a machine: the ultimate responsibility for AI/ML solutions relies with the senior management of the supervised institution', with clear roles along the whole AI life cycle.
What did CSSF White Paper on Artificial Intelligence (2018) change for banks?
This was the CSSF's first published study on AI. Its Annex 3 table of risks and recommendations (data, governance, ethics, technology, external providers) covers the trustworthiness themes, such as human oversight, bias, auditability and explainability, that the CSSF and the BCL later measured in their 2023 and 2025 thematic surveys. It remains non-binding.
Is the CSSF white paper on AI binding?
No. The CSSF says the white paper has no binding value vis-à-vis the supervised institutions. It is meant to provide the foundations for a constructive dialogue with the financial sector on the practical implementation of AI and its implications.
What does the CSSF white paper say about accountability for AI?
Its summary table states that accountability cannot be delegated to a machine: the ultimate responsibility for AI/ML solutions rests with the senior management of the supervised institution, which should define clear roles and responsibilities along the entire AI life cycle.
| Date | Document | Status |
|---|---|---|
| Jul 7, 2026 | CSSF communiqué on frontier AI cyber risk (7 July 2026) — Evolving opportunities and risks in artificial intelligence and its adoption | Final |
| May 16, 2025 | CSSF/BCL second thematic review on AI (2025) — Thematic review on the use of Artificial Intelligence in the Luxembourg financial sector (May 2025) | Final |
| Dec 23, 2024 | Luxembourg bill n° 8476 (AI Act implementation) — Projet de loi n° 8476 portant mise en oeuvre de certaines dispositions du règlement (UE) 2024/1689 (règlement sur l'intelligence artificielle) | Proposed |
| May 3, 2023 | CSSF/BCL thematic review on AI (2023) — Thematic review on the use of Artificial Intelligence in the Luxembourg Financial sector (May 2023) | Final |
| Nov 19, 2026 | BNM AI Feedback Statement — Feedback Statement: Artificial Intelligence (AI) in the Malaysian Financial Sector | Final |
| Nov 19, 2026 | BoE/PRA/FCA effective practices on AI cyber risk (Nov 2026) — Effective practices: Cyber risk and resilience management relating to artificial intelligence (AI) | Final |
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