Swiss AI Initiative: A Leap Forward in Transparent and Reliable AI Development
Artificial Intelligence & Machine Learning
Zurich, Switzerland – ETH Zurich and EPFL have launched the ambitious “Swiss AI Initiative,” setting a new benchmark in the global AI landscape. The initiative aims to establish Switzerland as a world leader in the development and implementation of transparent, reliable artificial intelligence (AI).
Alps Supercomputer: A Technological Marvel
February 2024 marks a significant milestone with ETH Zurich’s new Alps supercomputer going live at the Swiss National Supercomputing Centre (CSCS) in Lugano. Equipped with a staggering 10,000 graphics processing units (GPUs), Alps stands among the world’s most potent computing resources. Designed specifically for AI applications, this supercomputer empowers Swiss scientists with computing capabilities previously limited to major tech corporations.
Switzerland’s Competitive Edge in Digital Sovereignty
The Alps supercomputer is a strategic asset in Switzerland’s quest for digital sovereignty. “The rapid development of generative AI has created a global scarcity of supercomputing infrastructure, mostly controlled by large multinationals,” explains ETH Vice President for Research, Christian Wolfrum. “Our initiative seeks to leverage Switzerland’s location advantage, making AI expertise broadly accessible and ensuring independent research.”
Commitment to Transparency and Open Source
A core focus of the Swiss AI Initiative is the development of large language models (LLM) that are transparent, understandable, and adhere to legal, ethical, and scientific standards. “We emphasize transparency and Open Source, in stark contrast to most publicly available LLMs,” highlights Jan Hesthaven, Provost and VP for Academic Affairs at EPFL. Over the next year, the initiative will dedicate ten million GPU hours on the Alps computer for this purpose.
AI for Diverse Industries and Public Administration
This initiative will foster new AI base models for various sectors like robotics, medicine, and climate science, addressing key questions about human-AI interaction, ethics, security, and data privacy. Swisscom’s CTO, Gerd Niehage, sees the initiative as a crucial element in Switzerland’s digital future, accelerating digital transformation and generative AI capabilities.
Fostering Collaborations and Nurturing Start-ups
The Swiss AI Initiative aims to bridge science, industry, and politics, expanding partnerships across sectors. The initiative also focuses on supporting start-ups in the AI domain, ensuring that the benefits of this technological advancement permeate throughout the industry and public services.
Networking Swiss and International Researchers
ETH Zurich and EPFL, with their respective AI centers, will collaborate with the Swiss Data Science Center and other Swiss and international institutions. Already aligned with ELLIS, the European AI excellence network, this initiative underscores Switzerland’s commitment to pioneering multilingual, cross-border open source LLMs.
In conclusion, the Swiss AI Initiative stands as a testament to Switzerland’s vision of becoming a global AI hub, championing transparency, innovation, and collaboration in the artificial intelligence realm.
