Building on the successful first meeting in 2019, experts on AI recently met for a second time with not-for-profit organisations for the “AI for the Social Good” seminar at Schloss Dagstuhl in northern Saarland in Germany. The objective was to establish partnerships and build trust; to iterate on concrete problems in a hands-on hackathon; and to demonstrate what is feasible today.

Since 1990, Schloss Dagstuhl has been hosting computer scientists in what are known as the Dagstuhl Seminars. The center recently opened its doors to representatives of a completely different sector: development and humanitarian organisations. These organisations included AirQo - Makerere University from Uganda, D-tree from Tanzania, Humanitarian OpenStreetMap, Laterite, Netherlands Red Cross, Oxfam Novib, Save the Children, and TechnoServe. They came to challenge the AI experts from companies and various prestigious universities from around the world to help them with the global problems they are trying to solve.

Christoph Weigl from TechnoServe talked about the need for collaboration between the two fields and to demystify AI: "The exciting thing about this seminar is the opportunity for NGO practitioners and AI experts to test out where they have common interests, understand each other, and share the motivation to collaborate. At the same time, it is a reality check for everybody to share ideas and see whether and how artificial intelligence can be leveraged to address global challenges."

Equally enthusiastic was Nitusima Kataraia from D-tree: "What I like about this seminar is that I got a chance to exchange ideas with AI experts and receive their very valuable insights. Their suggestions will influence the direction of our Jamii ni Afya project. This project aims at improving the health of mothers and infants in Zanzibar. On top of that, I appreciate all the connections we made that will hopefully produce more fruitful results soon."

Stijn Koster from Save the Children explained why he saw value in attending the seminar. "NGOs produce massive amounts of data, but to make sense of it all, and to also put it to the benefit for others in the sector, still involves a lot of costly and time-consuming effort. Here at Dagstuhl, we can have meaningful interactions with academics who have know-how on AI's current possibilities to help alleviate this challenge."

Richard Sserunjogi from AirQo - Makerere University stressed the importance of the seminar. "The seminar provided an avenue for not-for-profit organisations to discuss some of the challenges they are currently facing with AI experts. These then helped us to provide concrete solutions and advice on how to use AI to maximise the impact of our data. Additionally, the seminar laid ground for further collaboration."

Experts in the AI field also see the benefit of teaming up. Nele Quast from the University of Oxford mentioned, "The workshop has offered a refreshing way to apply practical machine learning to real-world problems and feel that we can have an impact. It's been lovely meeting people with a wealth of experience with NGO and social impact work, along with machine-learning academics. Staying at Dagstuhl together meant we were able to get to know each other and hang out outside of working, too."

Mariella Goebl from TechnoServe concluded, "The Dagstuhl AI for the Social Good seminar is hands-on! It brings together AI experts looking for use cases and NGOs in need of solutions. This is the most fertile ground for long-term collaboration. I leave with a clear plan of what the next step is, and I know that we're taking this step as a team."

The AI for the Social Good seminar at the Schloss Dagstuhl Leibniz Center for Informatics had 22 people from 7 countries on 3 continents convene for a week to work on engendering collaboration. They braved the difficulties of travelling for work during this pandemic in order to live and work together on site at Schloss Dagstuhl for the week. The first two days focused on cracking each other's code.
Discussions focused on the NGOs' most pressing challenges and on what AI techniques are currently available to help solve them. On days three and four, the AI experts and the not-for-profit organisations held a hackathon to build concrete prototypes for AI applications for social good. The last day was used to plan for the future of the created collaboration.