A narrow, dated note on Swiss instruments relevant to AI-agent oversight. Switzerland has no comprehensive cross-sector AI statute; existing data-protection law and financial supervision carry the weight today, with an international convention being implemented on a multi-year timeline. Each entry names the instrument's exact status.
These pages describe the product and cite public instruments. They are not legal advice, not a certification, and not a statement that any instrument applies to your organisation. The official source is authoritative; applicability questions belong to your advisers.
Instruments, by status
| Instrument | Status | What it addresses |
|---|---|---|
| Revised Federal Act on Data Protection (FADP) | Binding law, in force 1 September 2023. The FDPIC has confirmed the FADP applies to AI systems, including automated individual decisions. | Personal data processing by AI systems; information and safeguard duties around automated individual decisions, subject to the FADP's own exceptions. Using AAES does not discharge the controller's duties; other parties' obligations depend on their roles. |
| Guidance 08/2024 on governance and risk management when using AI · FINMA (18 December 2024) | Supervisory guidance. Not legislation, but it states what FINMA expects in supervisory dialogue with supervised institutions; related instruments (for example Circular 2023/1 on operational risks and resilience for banks and securities firms, and the outsourcing rules) apply by institution category. | Central AI governance, a maintained AI inventory with risk classification, data quality, testing and monitoring, documentation, and responsibility that remains with the institution and cannot be delegated to a vendor. |
| Council of Europe Framework Convention on AI (CETS 225) · Swiss implementation | Signed, not yet ratified or implemented. The Federal Council decided on 12 February 2025 to seek ratification, and Switzerland signed on 27 March 2025; parliamentary approval, ratification and domestic implementation remain ahead. A consultation draft and implementation plan were announced for end-2026. Treat any "in force" claim for Switzerland as unverified. | Future duties on AI systems' conformity with human rights, democracy and rule of law; the Federal Council's stated approach is implementation through existing law where possible, with targeted amendments. |
| EU AI Act: extraterritorial exposure | Foreign law that can still apply. Swiss organisations placing AI systems on the EU market, or whose AI outputs are used in the EU, can fall in scope. See the European Union note. | Cross-border reach for Swiss-headquartered groups serving EU clients. |
What AAES records may contribute
| Expectation | AAES contribution | Capability status |
|---|---|---|
| Responsibility remains with the institution (FINMA Guidance 08/2024) | Every registered agent has a named accountable manager, a configured association rather than a statutory appointment; approvals are attributable to a named person; agent identities cannot act as approvers through the supported AAES approval path. | Implemented. |
| Human oversight of consequential actions (FADP automated-decision safeguards; FINMA expectations) | Actions the operator registers as irreversible (a registration classification, not an independent assessment of effects) require approval by an authorized person, bound to one intent and the exact payload, with expiry. An approval record does not by itself show the approver had adequate information, time or practical power to intervene. | Implemented. |
| Documentation a supervisor or internal audit can review (FINMA documentation expectations) | Sealed, hash-chained decision records with offline integrity verification by the client's own reviewer. | Implemented. An independently controlled witness can be configured separately; that is not an external assessment, and the default deployment has none. |
AAES decision records may support review of access and approval controls. They do not establish lawful processing, model fairness, suitability, or compliance with any instrument as a whole.
Within the configured trust model, integrity verification checks the exported records; it does not establish that every action was recorded, that an external action succeeded, or that recorded assertions are true. Approval records do not by themselves establish meaningful human oversight. Enforcement depends on control of the credential path, and actions bypassing AAES are not recorded.
Residual responsibility and explicit gaps
- The client remains the regulated entity. Whether the FADP, FINMA guidance or future convention-derived duties apply to you is your adviser's question, not this page's.
- Work outside AAES is invisible to AAES. Records cover actions routed through AAES; they say nothing about actions on bypass paths.
- No independent certification or assessment of AAES exists. No SOC 2 report, no penetration test.
- Status changes. The convention's Swiss ratification and implementation were still pending when this note was reviewed. Re-check the primary source before relying on the status shown here.
Bring the instruments your risk team cites to a scoped evaluation.
Related pages:
This note: https://aaes.ai/library/jurisdictions/switzerland.html
Hong Kong note: https://aaes.ai/library/jurisdictions/hong-kong.html
Singapore note: https://aaes.ai/library/jurisdictions/singapore.html
Trust model: https://aaes.ai/library/trust-model.html
Scope an evaluation: https://aaes.ai/contact.html?ref=library-evaluation