Stanford HAI, ETH Zurich, and EPFL – one of the world’s leading academic institutions in human-centered AI and engineering, have formed a strategic alliance at the World Economic Forum in Davos, Switzerland, to develop open-source, ethical foundation models as a transparent alternative to proprietary corporate AI. The partnership between Stanford and the newly formed Swiss National AI Institute (SNAI) aims to prioritize societal benefit and transparency over commercial profit.
The initiative builds on the success of Apertus, a high-performance open-source model trained in over 1,000 languages. By leveraging Switzerland’s “Alps” supercomputer located at the Swiss National Supercomputing Centre (CSCS) in Lugano, the alliance is set to bypass the massive “compute gap” that typically restricts such large-scale AI development to private corporations.
