The EU AI Act is not purely an EU matter: it can also affect Swiss companies. From 2 August 2026, large parts of the AI Act apply, including the transparency obligations for chatbots and AI-generated content. If you serve customers in the EU, now is the time to take a close look.
\nShortly before the deadline, the EU made adjustments: since 27 July 2026, the Digital Omnibus has been in force, postponing deadlines and redefining obligations. This does not mean that everything comes later, but that you need to know more precisely what applies when.
\nWe show when the EU AI Act applies to Swiss companies, what applies from August, what has been postponed and how you can prepare. We described the Swiss framework in our article on AI and data protection.
\nIn short: yes, if they offer AI in the EU or its output is used there.
\nUnder Article 2 of the AI Act, the regulation applies to providers that place AI systems on the market in the EU, wherever they are established. It also covers providers and deployers in third countries whose AI output is used in the Union. Even a chatbot answering enquiries from Germany can be enough.
\nYour role matters here: providers develop an AI system, or have it developed, and place it on the market under their own name; deployers use it professionally under their own authority. Many companies are deployers, for example with a chatbot they have bought in.
\nOn 12 February 2025, the Federal Council decided to ratify the Council of Europe’s AI Convention and to adapt Swiss law on a sector-specific basis wherever possible. A consultation draft on transparency, data protection, non-discrimination and supervision is to be prepared by the end of 2026 (Federal Office of Communications, OFCOM); the consultation has not yet been opened.
\nFor AI, the key law is the Swiss Federal Act on Data Protection (FADP). According to the Federal Data Protection and Information Commissioner (FDPIC), the FADP applies directly to AI-supported data processing. Where a language model communicates directly with users, they therefore have a right to know whether they are talking to a machine.
\nWhen the EU AI Act becomes generally applicable on 2 August 2026, Article 50 becomes central for many companies: the transparency obligations. Apart from a transitional period for older systems, they have not been postponed.
\nProviders must design chatbots and similar systems so that people are informed when they are interacting with an AI, unless this is obvious. The information must be given clearly, at the latest at the first interaction. Our article on chatbots for businesses shows how such systems work.
\nAnyone who uses AI to generate image, audio or video content that constitutes a deep fake must disclose that it has been artificially generated or manipulated. For evidently artistic or satirical works, a notice that does not hamper the work is sufficient.
\nAI-generated texts that inform the public on matters of public interest must also be labelled, unless they have undergone editorial review and someone holds editorial responsibility for them.
\nProviders of generative AI must mark the text, images, audio and video it generates in a machine-readable format; for systems on the market before 2 August 2026, the deadline is 2 December 2026. Help is available in a voluntary Code of Practice from the European Commission, published on 10 June 2026.
\nRegulation (EU) 2026/1744, officially the “Digital Omnibus on AI”, was published in the Official Journal of the EU on 24 July 2026 and has been in force since 27 July. Parliament and the Council had given their approval in June. The key deadlines:
\n| Rule | Applies from |
|---|---|
| Prohibited AI practices, AI literacy | 2 February 2025 |
| Transparency obligations under Article 50 | 2 August 2026 |
| Machine-readable marking for older systems | 2 December 2026 |
| New prohibitions: non-consensual intimate deep fakes, child sexual abuse material | 2 December 2026 |
| High-risk AI under Annex III, such as recruitment, creditworthiness | 2 December 2027 (previously 2 August 2026) |
| High-risk AI in products under Annex I | 2 August 2028 (previously 2 August 2027) |
The biggest postponement concerns high-risk systems, justified by delayed standards and the slow set-up of national authorities. Requirements such as risk management, documentation and human oversight do not disappear; they simply fall due later. If you use AI in recruitment, you gain time, but not a free pass.
\nPreviously, Article 4 required measures to ensure, “to their best extent”, a sufficient level of AI literacy. Since 27 July, providers and deployers must take measures to support it, without having to guarantee a specific level. Less obligation does not mean less need: untrained teams open the door to shadow AI.
\nWhat does this mean in practice? Five steps give you an overview, even if not everything has to be done by 2 August.
\nList which AI tools are used in your company, from chatbots to image generation. Tools that employees use on their own initiative also count.
\nFor each system, establish whether it reaches people in the EU or delivers results there. Typical examples are an online shop with customers in Germany or campaigns aimed at Austria.
\nAre you a provider or a deployer? If you offer an AI system under your own name, even one developed externally, you bear considerably more obligations than a company that uses an existing tool.
\nCheck whether your chatbots are clearly recognisable as AI, including those you have bought in. Set out in a short internal rule how AI-generated images, videos and texts are labelled.
\nAppoint a responsible unit, train your teams and document the measures. Our article on AI strategy shows how this fits into your overall planning.
\nFor Swiss companies, the EU AI Act is not a distant Brussels affair, but a question of market access. The Omnibus provides breathing space for high-risk systems, but hardly any for transparency. Switzerland, too, is moving towards transparency and supervision.
\nIt all starts with an overview. This is where BE BRAVE comes in: in our consulting, we assess AI readiness, data situation and risks and turn them into a prioritised roadmap rather than a list of ideas. All BE BRAVE offerings are operated on our own servers in Switzerland.
\nThe information reflects the situation as at 30 July 2026. This article is for general information only and does not constitute legal advice.
\nYes, if they offer AI systems in the EU or the output of their AI is used there.
\nThe transparency obligations apply from 2 August 2026. For the machine-readable marking of older systems, a transitional period runs until 2 December 2026.
\nNo. The Federal Council intends to ratify the Council of Europe’s AI Convention and is having a consultation draft prepared by the end of 2026. The FADP in particular already applies to AI today.
\nBreaches of the transparency obligations can lead to fines of up to EUR 15 million or 3 per cent of worldwide annual turnover, whichever is higher. For SMEs, the lower amount applies.
","featured_image":"/blog-media/eu-ai-act-schweiz/eu-ai-act-schweiz-header.jpg","category":null,"author_name":"BE BRAVE","tags":null,"is_featured":false,"status":"published","published_at":"2026-07-30T08:00:00.000000Z","seo_title":"EU AI Act Switzerland: rules from 2 August 2026","meta_description":"EU AI Act and Switzerland: what applies from 2 August 2026, what the Digital Omnibus postpones and how your company can prepare. Find out more here!","og_image":null,"canonical_url":"/en/eu-ai-act-schweiz","created_at":"2026-07-30T08:00:00.000000Z","updated_at":"2026-07-30T08:00:00.000000Z"}}