Stanford and Swiss Institutes Launch Global Alliance for Open Human-Centered AI

At the 2026 World Economic Forum in Davos, the Stanford Institute for Human-Centered AI signed a landmark agreement with the Swiss National AI Institute. This transatlantic partnership, involving ETH Zurich and EPFL, aims to develop open-source foundation models designed to serve the public good. By focusing on transparency and societal impact, the alliance seeks to re-establish academic leadership in a field currently dominated by private technology corporations.

Human-centered AI focuses on designing systems that enhance human capabilities and protect individual agency rather than simply automating tasks. This approach, often called Intelligence Augmentation, ensures that technical progress remains aligned with ethical standards and human values. Through this lens, the partners will create tools that empower researchers and citizens alike across various scientific and social disciplines.

Stanford HAI stands as a global leader in interdisciplinary research, exploring the ethical and policy implications of artificial intelligence. Its partner, the Swiss National AI Institute, represents a combined force of Switzerland's premier technical universities. Together, they bring a unique mix of social insight and high-performance engineering to the global stage, ensuring that AI development is not solely driven by commercial motives.

The collaboration will leverage significant technical resources, including the powerful Alps supercomputer housed at the Swiss National Supercomputing Centre. This hardware allows the academic alliance to train massive multimodal models that can compete with the proprietary systems of Silicon Valley. By using shared compute initiatives, the teams will focus on creating models that are multilingual and culturally diverse.

A key pillar of the project involves establishing rigorous evaluation frameworks and benchmarks for AI safety and reliability. These open-source tools will provide a neutral way for regulators and the public to assess the performance of large-scale models. By prioritizing open data and accessible architectures, the alliance ensures that the benefits of AI research are shared broadly across the global community.