Swiss Financial Market Supervisory Authority AI Guidance
FINMA expects supervised Swiss financial institutions to apply model risk management to AI: inventory AI applications, validate models, ensure explainability and human oversight, and keep data governance robust. Guidance 08/2024 sets out these supervisory expectations; they are principles-based rather than prescriptive rules, applied in proportion to each institution's AI use.
Updated 2026-09-06 · 19 tracked updates
FINMA provides regulatory guidance on the use of AI and machine learning in Swiss financial services. It covers model risk management, explainability, data governance, and supervisory expectations for banks and insurance companies using AI.
Who Needs to Comply?
Banks, insurance companies, securities dealers, and other financial institutions supervised by FINMA in Switzerland. Also relevant for fintech companies operating in the Swiss market.
Key Dates & Timeline
FINMA Guidance 08/24 on AI in financial services published 2024. Supervisory expectations for model risk management ongoing. Swiss AI regulation framework under development.
Upcoming Milestones
Planned Swiss AI bill target year-end 2026
Swiss FDPIC guidance on AI and data protection, plus legislative roadmapLatest FINMA Updates
Swiss FDPIC guidance on AI and data protection, plus legislative roadmap
The FDPIC has made clear that Switzerland’s data protection law already applies to AI-supported processing and that the federal government is targeting an AI bill by the end of 2026, so organizations should harden transparency and automated-decision controls now.
FINMA sets governance and risk-management expectations for AI use in Swiss financial institutions
FINMA’s 2024 guidance states that Swiss financial institutions using AI must identify, limit, control, and monitor AI-related risks within their existing supervisory framework, so firms should immediately test whether their governance and model-risk controls are adequate.
FDPIC says Swiss data-protection law applies directly to AI-supported processing
On 2025-05-08, the FDPIC reiterated that Switzerland’s data-protection law already applies directly to AI-supported processing, meaning transparency, purpose limitation, and data-subject rights must be built into current AI operations now.
FINMA guidance on governance and risk management when using artificial intelligence
FINMA’s 18 December 2024 guidance says supervised institutions must adapt governance and controls to the materiality and probability of AI risks, including operational, model, data, IT/cyber, third-party, legal, and reputational risks.
FDPIC AI and data protection guidance
The FDPIC states that Switzerland’s FADP applies directly to AI-supported processing and expects manufacturers, providers, and users to be transparent about purpose, functionality, and data sources.
FDPIC guidance on AI and data protection
The FDPIC’s AI guidance states that the Swiss FADP applies directly to AI-supported processing and requires transparency about purpose, functionality, and data sources, which elevates compliance expectations for AI deployments in Switzerland.
FINMA guidance on AI governance and risk management
FINMA’s AI guidance highlights operational, model, cyber, data-quality, third-party, legal, and reputational risks, so Swiss financial institutions should formalize AI governance and oversight now.
FINMA Guidance on Governance and Risk Management When Using Artificial Intelligence
FINMA published AI governance guidance on December 18, 2024, making governance, model risk, data quality, cyber risk, third-party dependence, and legal/reputational risk explicit supervisory priorities for Swiss financial institutions using AI.
Swiss FDPIC AI and data protection guidance
The FDPIC’s 2025 guidance makes clear that Swiss data protection law applies directly to AI-supported processing, so organizations must treat transparency, DPIAs, and human review rights as immediate design requirements for AI systems using personal data.
FDPIC confirms Swiss data protection law applies directly to AI processing
The FDPIC’s AI and data protection page confirms the nFADP has applied directly to AI-supported data processing since 1 September 2023, so Swiss organizations must treat AI use as already in scope for data protection controls.
Jurisdiction Coverage
Related Frameworks
Key Topics
Frequently Asked Questions
What does FINMA expect from financial institutions using AI?
FINMA expects supervised Swiss financial institutions to apply model risk management to AI: inventory AI applications, validate models, ensure explainability and human oversight, and keep data governance robust. Guidance 08/2024 sets out these supervisory expectations; they are principles-based rather than prescriptive rules, applied in proportion to each institution's AI use.
What does FINMA require for AI use in financial services?
FINMA expects supervised institutions to apply robust model risk management to AI systems, ensure adequate explainability and human oversight, maintain strong data governance, and conduct regular validation of AI models used in decision-making.
How does FINMA's approach differ from the EU AI Act?
FINMA takes a sector-specific, principles-based approach focused on financial services, while the EU AI Act is a horizontal regulation covering all sectors. FINMA's guidance is supervisory in nature rather than legislative, allowing more flexibility in implementation but expecting the same rigor in risk management.
Does Swiss AI regulation follow the EU AI Act?
Switzerland is not bound by the EU AI Act but is developing its own AI regulatory approach. Swiss companies serving EU clients must still comply with the EU AI Act. FINMA's guidance aligns with international standards while maintaining Switzerland's principles-based regulatory tradition.
Keep exploring
This hub tracks published FINMA guidance and related actions; it is not supervisory or legal advice, and expectations are applied institution by institution.