Translated with DeepL
Download discussion paper (in German)
In collaboration with its Artificial Intelligence Platform, the SATW published its recommendations for Switzerland's AI strategy in 2019.
The key points are still valid:
Artificial intelligence (AI) has long since found its way into our everyday lives. It is part of applications in our smartphones or those that are accessible via websites, such as real-time predictions of traffic congestion, powerful online translation services or, most recently, publicly accessible chatbots that provide information and suggest texts. Large technology companies are thus not only making their AI solutions available for their own services and products or industrial applications, but increasingly also for the public.
According to Anton Aschwanden, Head of Government Affairs and Public Policy at Google Switzerland, AI can make a significant contribution to the fulfilment of the UN's 17 sustainability goals and to mitigating other major challenges facing humanity. At the same time, the increasing use of AI also poses huge challenges. Responsible use of the technology is therefore crucial and indispensable.
This also means that beneficial, secure and trustworthy applications should be enabled and not unnecessarily hindered by regulatory provisions. According to Aschwanden, AI needs clear framework conditions that enable the broad and sensible use of the technology. In line with the statement by Google's CEO Sundar Pichai: "AI is too important not to regulate well".
With its world-renowned research institutions, innovative companies based here and liberal legislation, Switzerland is ideally placed to drive forward the positive development of AI. Switzerland therefore has many strengths. However, its federalist structures often make it difficult to keep up with the rapid pace of developments. Our country should therefore show more enthusiasm and determination, as control and access to this technology will also have an impact on its sovereignty, Aschwanden continued.
Data plays a central role in the context of AI and has great economic and social value. However, it must be of high quality and, above all, accessible. The recently adopted motion for a framework law for the secondary use of data emphasises that better access to strategically important data is needed to ensure security of supply. The scope and diversity of this data often requires AI in order to be processed in a useful way.
"We are realising the AI revolution today, but it started years ago," explained Behshad Behzadi, Vice President of Engineering at Google Switzerland, in the subsequent presentation. According to Behzadi, the generative AI revolution is now being added, made possible by large language models (LLMs) such as ChatGPT. Thanks to LLMs, it is almost impossible to distinguish between artificial and human language. There is also the time factor: it can be generated in a matter of seconds and simultaneously in different languages.
Google and many other providers make such solutions available to both business customers and private users via the cloud. Security, privacy, explainability and costs are key factors here. Many interfaces such as the Vertex AI API, the PaLM API or the generative language model Chirp are freely accessible. The responsible use of AI in accordance with the Google AI Principles is always important to Google.
Christoph Aeschlimann, CEO of Swisscom and SATW member, then addressed the question of how Switzerland's digital sovereignty can be increased. According to Aeschlimann, Swiss universities should focus on essential human problems. For example, Ecorobotix is developing a Swiss solution to a global problem: thanks to plant-specific spraying, pesticide-free agriculture seems within reach. If Switzerland produces companies as global market leaders in such relevant areas, it can become more independent. This is because such major players have a seat at the globally important tables and can get involved, as Roche and Novartis did in connection with the coronavirus pandemic, for example.
Switzerland undoubtedly has its strengths in established industries such as pharmaceuticals and banking. However, it has so far failed to produce global market leaders in important areas of the future. The problem is that there is too little venture capital available at an early stage for the growth of young companies.
At least CHF 11 billion in venture capital is needed to close the existing gap, said Aeschlimann. Otherwise, young companies would move abroad. Above all, start-ups in the decisive second investment round must remain in Switzerland. Because from a global perspective, companies that can scale faster than others will ultimately win. They generate jobs and retain talent. Corresponding investments could therefore make Switzerland a good deal more attractive and sovereign.
In the subsequent panel discussion, Swisscom CEO Christoph Aeschlimann, National Councillor Gerhard Andrey (co-founder and partner of Liip), National Councillor Judith Bellaiche (Managing Director of SWICO) and Alexander Ilic (Director of the ETH AI Centre) exchanged views.
Their tenor: Our society is not developing as quickly as AI technologies are. Euphoria about the new possibilities is not out of place. But we also need to critically scrutinise the impact of developments on our lives. Negative side effects such as deepfakes could lead to a defensive attitude among the population. This is understandable: fraudsters use AI to manipulate and distort photos, videos or audio files so deceptively real that they cause confusion or facilitate abuse.
According to the panellists, Switzerland has a good chance of playing a more active role in AI. AI is already needed today to operate many different systems and manage processes. Switzerland attracts top talent worldwide, but it is sometimes difficult to retain these specialists. By creating opportunities for talented people to establish themselves as founders, the pioneers working in Switzerland can make a more effective transition between research and commercialisation.
In view of the dependencies on foreign specialists, access to international markets and hardware, absolute sovereignty for Switzerland is neither possible nor sensible, according to the unanimous credo of the discussion panel. Resilience through networking is much more important: Switzerland must position itself as a good partner and focus on its own strengths: For example, playing up its neutral position in the development of reliable, trustworthy AI even more.
Google's motto regarding the further development of AI - bold, responsible, together - can also be applied to Switzerland: it should now create the conditions to enable the responsible use of AI and promote beneficial AI applications. After all, problems such as climate change cannot be tackled without AI. This requires participatory approaches and constant dialogue. The SATW can make a valuable contribution here with its network.
Manuel Kugler
Caspar Türler
