Luxembourg's Commission de Surveillance du Secteur Financier (CSSF) addresses AI through non-binding publications rather than AI-specific rules: a 2018 white paper on AI risks and recommendations, two joint thematic reviews with the Banque centrale du Luxembourg (May 2023 and May 2025) and, on 7 July 2026, a communiqué in which it 'expects all the members of your management body to establish governance structures that support effective management of frontier AI related risk' under DORA. The 2025 review, answered by 461 institutions, found 28% using AI in production or development and generative AI in 61% of reported use cases. A new CSSF/BCL AI questionnaire opened on 17 November 2026 and is due by 31 December 2026. Bill n° 8476, in committee since December 2024, would make the CSSF the AI Act market surveillance authority for the entities it supervises.
| Full name | 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 |
| Role | Integrated financial supervisor whose AI output is non-binding: a 2018 white paper, joint AI surveys with the central bank and a 2026 statement of frontier-AI cyber expectations under DORA; proposed AI Act market surveillance authority for its supervised entities |
| Force on banks | Supervisory guidance |
| Applies to | Entities supervised by the CSSF. The 2018 white paper states that it has no binding value; the 2023 and 2025 thematic reviews are survey reports; the July 2026 communiqué sets supervisory expectations for DORA entities under existing ICT risk-management rules. Bill n° 8476, not yet passed, would make the CSSF the AI Act market surveillance authority for AI systems placed on the market, put into service or used by the entities it supervises |
| Key document | Thematic review on the use of Artificial Intelligence in the Luxembourg financial sector (CSSF and BCL, 16 May 2025), read with the CSSF communiqué on frontier AI of 7 July 2026 |
| Latest move | 17 Nov 2026: the CSSF, jointly with the BCL, opened a new AI questionnaire on eDesk for banks, investment firms, payment and e-money institutions, specialised and support PFS, crypto-asset service providers and authorised fund managers, due by 31 December 2026; 7 Jul 2026: CSSF communiqué setting management-body expectations on frontier-AI cyber risk under DORA. |
| Documents tracked | 5 · all documents → |
The CSSF's first AI publication, the December 2018 white paper 'Artificial Intelligence: Opportunities, risks and recommendations for the financial sector', explains machine learning and its development life cycle, walks through financial use cases from chatbots and robo-advice to fraud detection and credit scoring, and ends with a table of risks and recommendations: human oversight depending on criticality, bias testing, explainability, audit trails, security by design and outsourcing controls. It states that it has no binding value for supervised institutions, and that 'Accountability cannot be delegated to a machine'.
Since then the CSSF has measured adoption with the BCL. The first thematic review (3 May 2023) covered banks, payment and e-money institutions surveyed in 2021 and early 2022: 30% used AI and AML and fraud detection led the reported use cases. The second (16 May 2025), extended to investment firms and fund managers, found generative AI in 61% of 402 reported use cases, 64% of respondents letting staff use public GenAI tools (60% of those without a dedicated policy), lower self-assessed auditability and explainability, and a weak grasp of the AI Act's risk categories. In its 2025 annual report (4 September 2026) the CSSF said it had contributed to the EBA's mapping of the AI Act's high-risk requirements for creditworthiness assessment and credit scoring against banking rules and to the ECB's work on banks' use of AI, and announced a third survey with results in Q1 2027.
Its sharpest statement so far is about cyber risk. The 7 July 2026 communiqué says frontier AI models could drastically shorten the window between a vulnerability's disclosure and its exploitation, reports that at many supervised entities vulnerability scans are not frequent enough and fixes come too late, and asks management bodies to govern frontier-AI risk under DORA, with measures such as exposure-based patching, containment engineering, a pre-built 'no-patch' response and AI for defence. On the AI Act, bill n° 8476 (filed 23 December 2024) would make the CNPD the default market surveillance authority, the CSSF the authority for AI systems placed on the market, put into service or used by entities it supervises, and the Commissariat aux assurances the authority for insurers; the Council of State's opinion of 10 July 2026 asks that the CSSF's competence be limited to AI directly linked to the provision of financial services.
What has the Luxembourg (CSSF / BCL) actually published on AI?
| 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 |
| Dec 21, 2018 | CSSF White Paper on Artificial Intelligence (2018) — Artificial Intelligence: Opportunities, risks and recommendations for the financial sector (CSSF white paper) | Final |
| Date | Type | Document / event |
|---|---|---|
| Nov 17, 2026 | Milestone | New CSSF/BCL questionnaire on artificial intelligence opens. As part of a joint initiative with the BCL, the CSSF launched a new iteration of its AI questionnaire for banks, investment firms, payment and e-money institutions, specialised and support PFS, crypto-asset service providers and authorised investment fund managers, to be submitted through eDesk by 31 December 2026. The CSSF's 2025 annual report said results would be published in Q1 2027. |
| Jul 10, 2026 | Milestone | Council of State gives its opinion on bill n° 8476. The Council of State's opinion on the AI Act implementation bill, adopted unanimously on 10 July 2026, notes that Article 74(6) of the AI Act limits the CSSF's competence to AI systems whose placing on the market, putting into service or use is directly linked to the provision of financial services, and asks, under threat of a formal objection, that the bill's CSSF provision say so. The opinion is in French. source ↗ |
| Jul 7, 2026 | Guidance | CSSF communiqué on frontier AI cyber risk (7 July 2026) — Evolving opportunities and risks in artificial intelligence and its adoption. In a communiqué of 7 July 2026 the CSSF warned that frontier AI models can increase the scale, speed and effectiveness of cyberattacks and drastically shorten the time between a vulnerability's disclosure and its exploitation. source ↗ |
| May 16, 2025 | Survey | CSSF/BCL second thematic review on AI (2025) — Thematic review on the use of Artificial Intelligence in the Luxembourg financial sector (May 2025). The CSSF and the Banque centrale du Luxembourg published their second thematic review on AI on 16 May 2025, based on a June to August 2024 survey answered by 461 financial institutions, an 86% participation rate. source ↗ |
| Dec 23, 2024 | Statute | 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). Bill n° 8476, filed in Luxembourg's Chambre des Députés on 23 December 2024, would implement the EU AI Act. source ↗ |
| Jun 19, 2024 | Milestone | CSSF and BCL launch the survey behind the second AI thematic review. The CSSF opened an eDesk questionnaire on the use of AI for banks, investment firms, payment institutions, e-money institutions and authorised investment fund managers, due by 7 August 2024. Its results were published on 16 May 2025. |
| May 3, 2023 | Survey | CSSF/BCL thematic review on AI (2023) — Thematic review on the use of Artificial Intelligence in the Luxembourg Financial sector (May 2023). On 3 May 2023 the CSSF and the Banque centrale du Luxembourg (BCL) published their first joint thematic review on the use of AI, based on a survey of credit institutions, payment institutions and e-money institutions run from October 2021 to January 2022. source ↗ |
| Dec 21, 2018 | Report | CSSF White Paper on Artificial Intelligence (2018) — Artificial Intelligence: Opportunities, risks and recommendations for the financial sector (CSSF white paper). 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. source ↗ |
- Results of the CSSF/BCL AI questionnaire open from 17 November to 31 December 2026, which the CSSF's 2025 annual report says will be published in Q1 2027.
- Passage of bill n° 8476 and the final wording of the CSSF's AI Act market-surveillance mandate after the Council of State's opinion of 10 July 2026, which asks that it be limited to AI systems directly linked to the provision of financial services.
- Supervisory follow-up to the 7 July 2026 frontier-AI communiqué, for example in ICT and DORA reviews of vulnerability management and management-body oversight.
- Any CSSF guidance on applying AI Act requirements in the financial sector: its 2025 annual report says it takes part in European working groups clarifying how the AI Act applies to financial entities.
Does the CSSF have binding rules on AI?
No AI-specific rules. The 2018 white paper says it has no binding value for supervised institutions, the 2023 and 2025 thematic reviews are survey reports, and the 7 July 2026 frontier-AI communiqué states supervisory expectations under DORA's existing ICT risk-management requirements. Binding AI obligations come from the EU AI Act; bill n° 8476 would make the CSSF its market surveillance authority for supervised entities.
How widely do Luxembourg financial firms use AI?
In the CSSF/BCL survey of mid-2024, published on 16 May 2025, 28% of 461 respondents used AI in production or development and 22% were experimenting or planning to within 12 months. Among banks, payment and e-money institutions, adoption rose to 43% from 30% in the previous survey. Generative AI featured in 61% of the 402 reported use cases.
What did the CSSF say about frontier AI and cybersecurity?
In a communiqué of 7 July 2026 it warned that frontier AI models can increase the scale, speed and effectiveness of cyberattacks and shrink the time between a vulnerability's disclosure and its exploitation. It expects management bodies to establish governance for frontier-AI risk under DORA and listed measures including reducing the attack surface, patching by exposure and exploitability, containment engineering, a pre-built no-patch response and using AI for defence.
Follow every move these regulators make
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