Experts on artificial intelligence (AI) from Google, Microsoft and academia are meeting with international not-for-profit organizations (NGOs) at Schloss Dagstuhl in northern Saarland in Germany this week. The question they seek to answer: can artificial intelligence benefit the social good?
Since 1990, Schloss Dagstuhl has hosted computer scientists at the world-renowned Dagstuhl Seminars. In a rare move, this week the center opened its doors to representatives of a completely different sector to collaborate with machine learning experts – staff from development and human rights organizations such as Oxfam, the International Commission of Jurists, RNW Media, Barefoot Law, SEMA, Justice & Peace, and Shaqodoon have come to challenge the computer scientists: what can you do to help poor farmers in Zimbabwe adapt to climate change? How can machine learning improve access to justice in Uganda? Is there a better way of dealing with toxic behavior in online debates?
“It is fascinating to finally work on real-world problems”, says Julia Proskurnia, a machine learning researcher at Google. “As researchers, we tend to look at problems from a very technical angle. Here at Dagstuhl we can have meaningful interactions with NGOs. We learn that our machine learning techniques can actually have a positive impact on the lives of people in poor countries.” Mustafa Othman from Shaqodoon in Somalia is equally enthusiast: “This seminar has definitely empowered me. I now see better how to put my data in a form that is useful for a developer. With any project, I will keep in mind how to collect data in a useful way.”
Gerald Abila from Barefoot Law in Uganda says that computer scientists were also made aware of particular challenges for NGOs: “During the seminar, we discussed with the computer scientists how we can ensure that their tools are designed so that they include vulnerable people and eliminate biases. Big data is the key towards understanding how to achieve the United Nations’ Sustainable Development Goals, and AI provides us with the means towards making sense of this big data. Artificial intelligence definitely has the potential to increase our impact. Dagstuhl offers a great opportunity to harness that potential.”
To make AI for social good really happen, Frank Mugisha from Rwanda stresses the importance of speaking the same language: “I liked the fact that we managed to achieve a common understanding between NGO people and AI people. We now understand from each other what needs to be done, and the way to do it.”
What next for AI for social good? For his organization in Somalia, Mustafa Othman is clear: “I am determined to make this collaboration with AI people work in the long run!”
This Dagstuhl Seminar at the Leibniz Center for Informatics has 24 people from 9 countries on 3 continents convene for a week to collaborate. The first two days, discussions focused on what the NGOs’ most pressing problems are, and on what techniques of artificial intelligence are available. On day three and four, the computer scientists and the not-for-profit organizations do a hackathon to build prototypes for AI applications.
The Dagstuhl Seminar is organized by:
- Claudia Clopath (Imperial College London, GB)
- Ruben De Winne (Oxfam Novib – The Hague, NL)
- Mohammad Emtiyaz Khan (RIKEN – Tokyo, JP)
- Tom Schaul (Google DeepMind – London, GB)
More information about the seminar 19082 – "AI for the Social Good" can be found at www.dagstuhl.de/19082