These three GenAI collaborations show the power of science diplomacy

Sometimes the channels of science remain open when understanding or negotiation is no longer possible at the political level. As a physicist, CERN¹ (the European Organization for Nuclear Research) comes to mind as an inspiring example of international scientific cooperation. Founded in the middle of the Cold War, CERN has served as neutral ground in a time of political polarization.

In recent years, the Synchrotron-Light for Experimental Science and Applications in the Middle East (SESAME)² demonstrates another example. Operating in Jordan, researchers from Israel, Palestine, and Iran, among others, work together. Switzerland has also increased its efforts in the field of science diplomacy, including through the Geneva Science and Diplomacy Anticipator (GESDA),³ founded in 2019, in which researchers from ETH Zurich remain actively involved.

These examples illustrate how science diplomacy can transcend the geopolitical arena, building trust and bridging gaps between nations.

Swiss science diplomacy

Raised in Geneva, science and diplomacy have become an integral part of my life experience and the daily business of my role as a university president. While diplomacy often addresses current global concerns, science is anticipating future challenges.

Switzerland’s reputation for neutrality places it in a unique position to facilitate global cooperation in science and technology, innovation, and knowledge exchange. As a trusted intermediary, the country fosters partnerships across industry, academia, and intergovernmental agencies. 

To illustrate, consider these three examples of science diplomacy at work. These collaborations in generative artificial intelligence (GenAI) highlight the power of cross-sector cooperation in shaping a better future.  

Multilateral moves towards open-source AI

Ubiquitous in our everyday lives, we’ve been using artificial intelligence (AI) for more than a decade and have hardly noticed. Whenever you consult your digital assistant, your wellness wearable, or you pass a human authentication test you become an AI consumer.

With the arrival of ChatGPT and GenAI models, the way we work and how we produce is transforming at an overwhelming pace. While there is no doubt about the significant potential of large language models (LLMs), GenAI comes with several risks and imperfections which must be dealt with.

Computer scientists at Switzerland’s two federal universities, ETH Zurich and EPFL, working with the Bulgarian AI research institute, INSAIT,⁴ and ETH Spinoff, LatticeFlow AI,⁵ put GenAI to the test to see if it meets the European Union (EU) AI Act’s mandate for transparent and trustworthy AI. The research team developed the first comprehensive technical interpretation of the EU AI Act, translating the act into measurable technical requirements for AI and assessing how well AI models comply with legal requirements.

Meanwhile, academia and industry partners such as Google and IBM are working on frameworks for open innovation and the development of GenAI as a potential tech for good, while minimizing the real possibility for harmful AI. As a member of IBM’s AI Alliance,⁶ ETH Zurich plans to host a series of multi-stakeholder roundtables to address potential roadblocks that hinder the democratization of AI on a global scale. They will also examine the challenges and opportunities associated with the implementation of the EU AI Act.   

Helping the Red Cross dispatch supplies

ETH Zurich researchers developed a planning tool for the International Committee of the Red Cross (ICRC) that made the extraordinarily complex task of providing critical medical supplies to people in need more efficient. The tool was deployed in a dozen locations in Africa, the Middle East, and Ukraine and, in 2023, saved the ICRC an estimated CHF3.6 million (approximately 4.1 million USD).

Now, a new project will explore how the AI tools used in commercial supply chains of companies like Amazon could be adopted by humanitarian organizations, where resources are scarce and lives are at stake. The project will investigate the potential of machine-learning to improve how the ICRC forecasts the delivery of medical supplies to health facilities in war zones. It will also consider how to apply AI in building a predictive model to improve the accuracy of future forecasts.

Supporting UN peacekeeping missions

Researchers from the Center of Security Studies (CSS)⁷ and Interactive Visualization & Intelligence Augmentation Lab (IVIA)⁸ at ETH Zurich are working with the United Nations (UN) to further develop a deep-learning model to predict short-term conflict events using techniques adapted from computer vision to support UN peacekeeping missions.

The long-term vision is to create a collaborative human-AI platform involving deep neural networks and a diverse pool of human experts, who can jointly make event forecasts and informed decisions for conflict zones. The project is just one element of a recent partnership forged between the UN and ETH Zurich.

The above initiatives in GenAI, humanitarian aid, and peacekeeping demonstrate the transformative power of cross-sector collaboration to realize meaningful progress. It helps us to address challenges more effectively in our rapidly changing world. Growing global collaboration through science and technology is now of the utmost importance.

Overcoming political and societal barriers to innovation by promoting open and multilateral cooperation is vital. In this regard, Switzerland, as a neutral broker, can play an important role in science diplomacy and help to ensure that progress serves the greater good and contributes to a more connected and resilient global society.

Joël Mesot has written this post for the external page WEF Agenda Blog, where it first appeared.