“Is there actually an AI law in Switzerland yet?” It’s a valid question. The EU has its AI Act, the media is talking about deadlines and fines, and many executives feel like they’re missing out on something.

The short answer: No, Switzerland does not yet have its own AI law. But that doesn’t mean AI operates in a legal vacuum. Many rules have long been in place, and the Federal Council is preparing the next phase. This article breaks down what’s currently in effect, what’s coming and when, and what you can do right now in your company.

Is there an AI law in Switzerland?

No, as of October 2026, there is no specific AI law. Switzerland has deliberately chosen not to adopt a broad, horizontal law modeled after the EU’s. Instead, existing laws are being amended where AI raises new issues, and the Council of Europe’s AI Convention is to be incorporated into Swiss law.

What this means for you is that anyone using AI today must operate within a framework of data protection laws, fair trade laws, copyright law, liability law, and industry-specific regulations. These rules are formulated in a technology-neutral manner. They do not ask whether a human or a machine is making the decision, but rather what happens to data, customers, and the rights of third parties.

What the Federal Council Has Decided

On February 12, 2025, the Federal Council established its regulatory approach based on an analysis by DETEC and the FDFA. Three decisions have shaped the path since then:

  1. The Council of Europe’s AI Convention is to be ratified. The Council of Europe’s AI Convention is the first legally binding international agreement on AI and requires that AI systems respect human rights, democracy, and the rule of law. It is primarily aimed at government actors. Federal Councilor Albert Rösti signed it on March 27, 2025, in Strasbourg.
  2. Adjustments should be made on a sector-by-sector basis whenever possible. Instead of a single AI law covering everything, the relevant laws will be amended in a targeted manner. Overarching rules should apply only where fundamental rights are centrally affected, such as in the area of data protection.
  3. Supplementary, non-binding measures. Self-declarations and industry-specific solutions are intended to supplement the legislation.

The Federal Council has clearly assigned the tasks for this. The Federal Department of Justice and Police (FDJP), together with the Federal Department of the Environment, Transport, Energy and Communications (DETEC) and the Federal Department of Foreign Affairs (FDFA), will draft a consultation document by the end of 2026. It is intended primarily to regulate transparency, data protection, non-discrimination, and oversight. At the same time, DETEC—led by OFCOM—is working with the FDJP, the FDFA, and the EAER to develop an implementation plan for the non-legally binding measures by the end of 2026.

In a nutshell: The general direction has been set, but the details have not. Only the draft bill will reveal how strict the transparency requirements will be and who will be responsible for oversight. In May 2026, the Federal Council stated that it was still too early to say; the goal is a level of protection equivalent to that of the EU.

Schedule: What Happens and When

The rule here is: Legislation in Switzerland takes time, and that is by design. Each step gives the business community, the cantons, and associations the opportunity to have a say.

  • February 12, 2025: The Federal Council makes a policy decision to ratify the AI Convention.
  • March 27, 2025: Switzerland signs the AI Convention.
  • End of 2026: The FDJP will submit the bill to the Federal Council; the implementation plan for non-binding measures is also expected to be ready by then. As of early October 2026, the consultation process has not yet begun.
  • First half of 2027: Opening of the public consultation period, as planned by the Federal Council, with a target date of late June 2027. It generally lasts at least three months.
  • Next: Evaluation, Federal Council message, deliberations in the National Council and Council of States.
  • Effective Date: Based on our assessment and the usual process, this will take effect no sooner than a few years after the consultation period begins.

We record every new step in our consultation tracker.

What does this mean for you, specifically? You have time to prepare. But you should make the most of it, because the draft will be based on the AI Convention and the EU’s experience.

What Already Applies Today

There is no law with “AI” in the title, but the rules behind it still apply. It’s like traffic laws: For a long time, there were no specific regulations for e-scooters, but the traffic rules still applied.

  • Data Protection Act (DSG): The revised Data Protection Act (DSG) has been in effect since September 2023 and is technology-neutral. In May 2025, the EDÖB reaffirmed that it applies directly to AI applications. Anyone who develops, offers, or uses AI systems must provide transparency regarding the purpose, functionality, and data sources. Where there are high risks to data subjects, the law requires a data protection impact assessment. The article “Using AI Safely” explains how to enable ChatGPT, Copilot, and other cloud-based AI in compliance with data protection regulations.
  • Unfair Competition Act (UWG): Anyone who misleads others using AI-generated content—such as fake reviews or misleading advertising—is acting unfairly under the Unfair Competition Act (UWG). There is no general labeling requirement for AI-generated content in Switzerland. However, the FDPIC infers from the DSG that users must be informed when they are speaking to a machine, and that fake images or voices of identifiable individuals must be clearly recognizable.
  • Copyright Law (URG): Anyone who uses AI-generated results must respect the rights of third parties under copyright law (URG). Only human creations are protected; according to the Institute for Intellectual Property, anything generated by AI in response to a routine command is not protected. The main point of contention is whether training an AI using protected works constitutes a copyright infringement.
  • Liability: The general provisions of the Swiss Code of Obligations and the Product Liability Act apply. There is no AI-specific liability law; a motion to modernize the Product Liability Act with regard to AI is pending, and the Federal Council supports it.
  • Industry Guidelines: In Supervisory Circular 08/2024, FINMA outlined what it expects from all supervised entities—such as banks and insurance companies—regarding AI governance and risk management.

It’s important to note: These rules apply regardless of whether you use ChatGPT, Copilot, or a Swiss model. The impartial comparison of Swiss AI chatbots shows which providers process data and where.

And what about the EU?

What many people overlook: For many Swiss companies, the EU AI Act is more important today than any future Swiss law. It also applies to companies outside the EU as soon as they offer AI systems in the EU or the results are used there. A chatbot on a website for customers in Germany is enough to trigger this.

Our guide to the EU AI Act for Swiss companies explains which obligations have been in effect since August 2, 2026, what has been postponed under the Digital Omnibus, and how you can check whether you are affected.

Copyright and Deepfakes: What Parliament Is Discussing

In addition to this major bill, there are two political debates you should be aware of.

Copyright. The Gössi motion (24.4596) calls for better protection of journalistic content and other copyrighted works against use by AI providers. The National Council adopted it in September 2025 with an amended text, and the Council of States approved it in December 2025. The text calls for comprehensive protection without undermining Switzerland’s position as a hub for business and innovation; how this will be implemented remains to be seen. The IPI is examining several models, such as a remuneration-based limitation or an opt-out mechanism that would allow authors to object to the use of their works. The preliminary draft is scheduled to enter the consultation process at the end of 2026. swissAI has submitted comments on the matter.

Deepfakes. In May 2025, the National Council rejected a general regulation on deepfakes. The Federal Council’s reasoning: Criminal and civil law already cover such misuse. The debate is not over yet: In June 2026, the National Council adopted a motion calling for a risk assessment of generative AI capable of producing sexualized deepfakes, and according to the Federal Council, measures against deepfakes are being examined as part of the AI bill. Read more in our article on sexualized deepfakes.

What Companies Should Do Now

You don't have to wait for the law to be enacted to be prepared. The following steps are worthwhile in any case because they comply with current law and prepare you for the upcoming bill.

  1. Track AI usage. Compile a list of which AI tools are being used in the company, by whom, and with what data. Many companies are surprised by how much is already in use.
  2. Review data protection measures. For each tool, determine where the data is processed, whether the provider uses it for training purposes, and whether a data processing agreement is in place. If there are high risks: Conduct a data protection impact assessment.
  3. Determine whether the EU applies. If you have customers, employees, or products in the EU, review the AI Act.
  4. Establish rules. A brief AI policy provides clarity: what is allowed, what isn't, and who makes the decisions. The AI Strategy Package helps you get started.
  5. Train employees. Anyone who uses AI should be aware of its limitations. That’s the best way to prevent errors anyway. You can find relevant courses in the overview of AI training programs.
  6. Communicate transparently. Let customers know when they are speaking with a chatbot or when content is generated by a machine—in situations where they might be misled.
  7. Have a say. The consultation process is an opportunity to share practical experience. swissAI will submit comments and incorporate its members’ experiences.

Where swissAI Stands

swissAI welcomed the signing of the AI Convention and called for its swift implementation, including an education initiative. At the same time, the association warns against obstacles to innovation: Regulation should limit risks without hindering its application in small and medium-sized enterprises. As the sponsoring organization of the Parliamentary Group on Artificial Intelligence, swissAI brings practical knowledge directly to the Federal Parliament.

Frequently Asked Questions

Is there an AI law in Switzerland?

No. As of October 2026, there is no specific AI law. The Data Protection Act, the Unfair Competition Act, copyright law, liability law, and industry regulations such as FINMA Supervisory Circular 08/2024 apply.

When will the AI Act be enacted?

The FDJP will submit the bill to the Federal Council by the end of 2026; the consultation process is scheduled for the first half of 2027. This will be followed by the government message and parliamentary deliberations. According to the usual procedure, it will take several years for the new rules to take effect.

Does the EU AI Act apply in Switzerland?

This does not apply as Swiss law, but it does apply to Swiss companies with ties to the EU. Anyone who offers AI systems in the EU or uses their results there must comply with the obligations set forth in the AI Act.

Do I have to label AI-generated content in Switzerland?

There is no general labeling requirement. However, misleading practices are prohibited under fair trading laws, and the DSG requires transparency when personal data is processed using AI. For content published in the EU, the transparency requirements of the AI Act apply.

Conclusion

Switzerland is deliberately regulating AI in a phased manner: adopting the AI Convention, making targeted adjustments to existing laws, and promoting industry-specific solutions. For companies, this does not mean sitting on the sidelines. Current laws—most notably the Data Protection Act—are already in effect, and any company operating in the EU is already subject to the AI Act.

In a nutshell:

  • Switzerland does not yet have its own AI law.
  • On February 12, 2025, the Federal Council decided to ratify the Council of Europe’s AI Convention and to amend laws on a sector-by-sector basis wherever possible.
  • The FDJP will submit the draft legislation on transparency, data protection, non-discrimination, and oversight by the end of 2026; the public comment period is scheduled for the first half of 2027.
  • The revised DSG applies directly to AI applications.
  • The EU AI Act already affects Swiss companies with ties to the EU.
  • An inventory, a data protection audit, an AI policy, and training are all worthwhile investments right now.

How does your company handle AI policies? Join the discussion on Friday Weekly or get started with free access to the community.

As of October 6, 2026. This article is intended as a general overview and does not constitute legal advice; the assessment depends on the specific circumstances of each case.

Sources: Federal Council on the Regulatory Approach, February 12, 2025 · Federal Council on the Signing of the AI Convention · Federal Office of Justice, Artificial Intelligence · OFCOM, Artificial Intelligence · FDPIC on the Data Protection Act and AI · IPI, FAQs on AI and Copyright · Curia Vista, Gössi Motion