SWITZERLAND Law and Practice Contributed by: Dirk Spacek, Wenger Plattner
• Civil liability – the CO and CC govern liability for damages caused by AI. • Sectoral regulations – AI in healthcare, finance or transport may be subject to sector-specific regula - tory frameworks. The Swiss Federal Council published a strategy paper on AI (2020) promoting innovation while emphasis - ing ethics, transparency, and human oversight, but legally binding AI-specific obligations are still limited. In March 2025, the Swiss Federal Council joined and ratified the Council of Europe’s AI Framework Conven - tion, and committed to adapt standards of this con - vention into Swiss law accordingly over time. While this ratification has not yet taken effect, it will primarily apply to state actors at the federal level and would only in exceptional cases have a horizontal effect on private parties (eg, where fundamental rights are at stake and must be reflected in contracts with govern - mental bodies). 3.1.2 AI-Related Lower-Level Regulation While Switzerland has no general AI law, sector-spe - cific regulations can impose requirements on AI use. • Healthcare – AI in medical devices is regulated, eg, under the Medical Devices Ordinance (MedDO, SR 812.213) and Swissmedic guidelines, requiring safety, performance and risk management. • Finance – financial institutions using AI must com - ply with various FINMA regulations on risk man - agement, transparency and model validation. • Transport and safety-critical systems – AI in autonomous vehicles or aviation is subject to Fed - eral Roads Office (FEDRO) and Swiss Civil Aviation Authority (FOCA) standards. • Data protection and ethics – all sectors must comply with the FADP and should adhere to ethical principles for fairness, transparency and account - ability, as recommended by the Federal Council’s AI strategy. 3.1.3 AI-Related Self-Regulation Many Swiss companies self-regulate AI development and use due to the absence of comprehensive AI- specific legislation. Self-regulation typically involves internal policies and guidelines, ethical AI principles, risk management frameworks, and human oversight
mechanisms, data governance (FADP), standards and best practices, adoption of international standards (eg, ISO/IEC 42001 on AI management) and secto - ral guidelines, and internal audit to review algorithmic performance, bias and security. 3.2 Contractual Requirements With Respect to AI 3.2.1 Key Requirements Sought by Customers When contracting AI solutions, Swiss customers typi - cally seek assurances covering risk, performance and compliance. They mostly focus on transparency and explainability, compliance with data protection, per - formance and accuracy, liability and indemnifications for errors, harm, or regulatory breaches caused by AI outputs, audit and monitoring rights, and ethical and regulatory compliance. Contracts often combine technical, legal and ethical safeguards to manage the inherent risks of AI solutions. 3.2.2 AI-Related Liabilities in Contracts In Switzerland, AI liability is primarily addressed contractually, since no AI-specific law exists. Key approaches include the following. • Allocation of responsibility – contracts specify whether the supplier or customer bears risk for errors, biased outputs, or harm caused by AI. • Warranties and performance guarantees – sup - pliers often warrant compliance with contractual specifications, accuracy thresholds, and regulatory obligations. • Indemnification clauses – suppliers may agree to indemnify customers for damages, regulatory fines or third-party claims resulting from AI misuse or failures. • Limitations of liability – parties often cap financial exposure, exclude indirect damages, and define triggers for liability. • Audit and monitoring rights – customer over - sight provisions mitigate risk and support liability enforcement. 3.3 Key Concerns of Providers Providers of AI solutions in Switzerland face several legal, operational and reputational concerns.
80 CHAMBERS.COM
Powered by FlippingBook