Switzerland - Market Insights

In a time where businesses rely heavily on their technological portfolio to remain competitive, substantiate their valuation or position themselves validly as disruptive market entrants, AI has come to play a central role. In fact, few businesses and industries now seem willing to forgo references to AI in one way or another.

AI itself is not new. Going back to Alan Turing and the computing developments of the 1950s, it was already all the rage in the 1980s — though this author was not old enough to testify to this — before fading away for nearly three decades (albeit for the ludic AI victories of IBM’s Deep Blue against Garry Kasparov at chess in 1997 and DeepMind’s AlphaGo against Lee Sedol at Go in 2016 (Lee Sedol’s later statement upon retiring in 2019 is sobering: “With the emergence of artificial intelligence, being number one no longer has meaning”)). Then came the watershed moment of 30 November 2022 when OpenAI’s ChatGPT came crashing into the daily lives of the masses.

Apart from the mind-blowing technological advances, what is new is the quasi-infinite use cases of AI in its current forms. These use cases play out on different axes: (i) from the legal to the potentially illegal (for instance, the disturbingly problematic deepfakes); and (ii) from high-tech problem solving (e.g. DeepMind’s AlphaFold’s 3D protein structure prediction, autonomous vehicle driving, and so forth) to useless or simply unnerving (e.g. automated Tweet drafting, generated product review blurbs, AI hotline assistants, and numerous so-called “GPT wrappers”).

To avoid the fastidious exercise of defining all key terms pertaining to AI (e.g. machine learning, neural networks, large language models, even AI itself), we use the term “AI” in the most generic layman’s sense. This does have its limitations, however. Indeed, although markets around the globe are rife with businesses boasting new “AI” tools, the crushing majority are mislabelled — if not simply misleading — as the infamous “powered by AI” tag suggests. The reality is boringly binary: in terms of AI developments (ownership of the technology and innovation, computing power, funding and revenues) there is, on the one hand, the US and China and, on the other hand, the rest of the world. And it is not even close.

This gap has paradoxical implications for European and Swiss market actors. Consumers and businesses rely, with great appetite, quasi-exclusively on American or Chinese AI technology, yet want the comfort of some form of sovereignty and protection. Governments financially incentivise research and development (R&D) in the AI spheres, yet parliaments heavily regulate AI, thereby effectively elevating hurdles to the market deployment of the results of such incentivised R&D. As a result, there is a pervasive feeling among market actors that we set the rules, but others play the game.

The question of whether and how to regulate AI is, in Switzerland and, supposedly, across the European continent, the most relevant trend at the moment. More specifically, the issue is where do we — citizens, consumers, businesses, lawyers, society at large — wish to place the cursor on regulatory intervention.

The European Union’s pioneering AI Act — a comprehensive regulation of 2024, the requirements of which are gradually entering into effect over time — seeks to ensure the safety of AI systems, protect fundamental rights across the EU, provide a harmonised and predictable legal environment and foster innovation. It is the torchbearer for the systematic and extensive regulation of AI and carries with it the hopes of placing the EU at the centre of AI innovation while promoting values of trust and safety. It is also the lightning rod for criticisms; and such criticisms are rather easy, as its various critics may gladly confirm. Indeed, it would seem rather straightforward to build the case that the EU’s AI Act — in conjunction with numerous other regulations and directives that the EU Parliament has issued over the past 10 to 15 years (e.g. the General Data Protection Regulation (GDPR), Data Act, Digital Markets Act (DMA), Digital Services Act (DSA), to name but a few) — has had a detrimental effect on innovation. This is, the argument goes, because the more regulations there are, the more arduous, costly and time-consuming the compliance and ultimately the market entry and deployment become. Numerous regulations lead to a fragmented regulatory landscape and numerous, sometimes unforeseeable, hurdles towards compliance. It is not much of a stretch, then, to argue that instead of providing local market actors with a safe and sound regulatory background for the development of tomorrow’s AI technology, such regulations have further cemented the central standing of the legacy technology companies (often US-based) and left a generous open space for other national legislators who may have more willingness towards a softer regulatory approach (here primarily China).

The above arguments are valid to a certain extent and this author’s practice tends to confirm, albeit anecdotally, that local market actors are often more comfortable expanding on North American and Asian markets rather than European ones due to concerns around EU regulatory requirements. More assertively, an October 2025 study requested by the European Parliament’s Committee on Industry, Research and Energy (ITRE) and titled “Interplay between the AI Act and the EU digital legislative framework” notes the highly burdensome legislative landscape, the fragmented state of EU digital legislation, the disproportionate compliance burden on European AI innovators and the overall detrimental effect on smaller companies compared to multinational enterprises. A case in point is the European Commission’s current efforts to simplify and transversally align requirements across various laws, including the AI Act, in its Digital Omnibus proposal.

We should not however lose sight of the necessity of a base layer of legislation on this important topic, as well as the need to define clear boundaries to adequately channel innovation in the AI realm. The EU’s approach deserves much credit in this respect. In addition, it is a common mistake (at least for lawyers) to over-emphasise the importance of legislation in the success — or lack thereof — of innovation; indeed, the legislative and regulatory landscape is only one among many factors enabling — or slowing down — innovation in any given market.

Perhaps, therefore, it is a question of “when” rather than a question of “whether” to legislate. This has certainly been the thought process that Switzerland has been following. Though not part of the EU, Switzerland is on many levels closely intertwined with it. In a country of roughly nine million people, Swiss market actors quickly look towards EU markets. In turn, the Swiss Parliament often seeks to align internal legislation with EU requirements, if only to facilitate trade with our neighbours. This means that Switzerland is rarely the first to legislate in any given area. The Swiss legislator has also historically been rather cautious with the intensity of its prescriptive legislation, frequently relying on general (i.e. not detailed) overarching laws, prudential rules or soft law. With very few exceptions, Swiss laws are also steadfastly technologically neutral, thereby ensuring (in theory) their resilience over time.

Accordingly, Switzerland does not have a general AI law and is unlikely to adopt in the near future a close adaptation of the EU’s AI Act (contrary, for instance, to what Switzerland did in the area of data protection). Rather, the Federal Council — Switzerland’s executive — opted for: (i) ratification of the Council of Europe’s AI Convention, a treaty establishing core obligations for AI systems to respect human rights, democracy and the rule of law (primarily targeting state actors); and (ii) limited sector-specific adjustments to various existing laws (such as in the health and transport sectors). This remains a work in progress from a legislative standpoint, however, as the Swiss government is reviewing other necessary changes and the implementation of non-binding measures. More generally, Switzerland remains committed to providing an innovation- and business-friendly environment, while protecting society at large and ensuring legislative compatibility with Switzerland’s primary business partners. Interestingly, this approach is at the same time cautious — in effect maintaining a status quo in terms of legislative action — and also a delicate balancing act between broadly diverse interests (on the internal market and international levels).

The Swiss approach also has its strengths and weaknesses. In a rapidly evolving technological landscape, a lighter legislative approach enjoys strong support from businesses. In recent years, this was, for instance, the case with Switzerland’s baseline DLT (distributed ledger technology, blockchain) Act. This is all the more true in today’s digitalised world in which market actors can move from one market to another and therefore may opportunistically choose what they consider as the most suitable environment. On the other hand, the Swiss approach is typically slow by international comparison and the laws risk having little relevance by the time they enter into force.

From a practitioner’s perspective, the Swiss legislative landscape is conducive to AI innovation and also adequate to protect the interests of society, consumers and citizens, against the risks of AI. This is, however, perceived as a temporary solution until further technological advances arise, as they always do. As mentioned earlier, international expansion is where Swiss businesses usually face the strongest friction in terms of compliance efforts. Perhaps in part for that reason, numerous Swiss AI-focused businesses quickly look to straddle the Swiss and North American markets before seriously looking to expand across other European markets. This dynamic is expected to change once the EU regulatory landscape becomes more familiar and maybe further streamlined, bearing in mind also the further incentives in place in Switzerland and across Europe to bolster AI innovation and offer a fertile ground for the growth of AI companies.

All this goes to illustrate the challenges of legislative and policy actions when it comes to new technologies. We live in a fascinating time and legislator and policy-maker savviness will play an outsized role in the coming years.