Better Breakthroughs: Four calls to action for the AI for Good community
What do we need to build a better and more equal world for everyone on this planet? And how can artificial intelligence (AI) help us get there faster?
The AI for Good Global Summit organized annually by ITU and XPRIZE Foundation in partnership with 37 UN agencies and ACM, and co-convened with Switzerland, aims to find practical applications of AI to advance progress towards the UN Sustainable Development Goals (SDGs).
This year, the main mission of the Breakthrough Days event was to mobilize the power of AI in advancing progress across three domains: gender equity, food security and dealing with pandemics.
Teams from around the world submitted their projects addressing one of the three domains and nine teams (the top three in each category) were selected to present at the Breakthrough Days event. Such projects included using AI to stem the spread of COVID-19, improving satellite imagery with AI to increase agricultural yield, developing a set of guidelines to inform judiciaries on how to address gender issues in the design and implementation of AI systems, and more.
Drawing on the lessons learned through the course of Breakthrough Days, featuring keynote speeches and interactive workshops, four initiatives directly addressing the challenges and opportunities of using AI in these three areas were announced during the closing session.
Call to action 1: Sign the Global Data Pledge
Data is the key component for artificial intelligence to become an effective tool in resolving global problems - but access has been a long-running obstacle. However, as Yoshua Bengio noted in his earlier Breakthrough Days keynote, data is often expensive or kept confidential due to proprietary issues.
During a crisis, however, a lack of data at hand can thwart much-needed progress.
This is why the Global Data Pledge is asking national governments, institutions and private organizations to sign a pledge to share data when the need arises.
"Having the right data at the right moment completely changes our understanding of a crisis, whether it's a pandemic, a tsunami or an environmental disaster," said Richard St. Pierre, CEO of C2 International, while making the announcement.
St. Pierre clarified that the Pledge will not ask organizations to reveal private information. Rather, the organizations will commit to sharing data and collaborating with each other under their own terms during an emergency.
The Pledge itself is an outcome of the Global Initiative on AI and Data Commons set up by XPRIZE and ITU following the 2019 AI for Good Global Summit.
St. Pierre also acknowledged concerns over privacy and data storage, calling the Pledge “a first step” in removing barriers to data sharing. Rather than replacing ongoing open data initiatives, he said: "What we are aiming for is to act as a catalyst for the exchange of that data to take place."
How to participate: The pledging will officially begin in January 2021, but St. Pierre has invited those who are interested or who want to help design the playbook of data pledges to join their LinkedIn working group or reach out to him by email.
Call to action 2: Help develop an inclusive framework for the future of food
The growing problem of food insecurity is not just about solving hunger. Food systems are closely interlinked with sociological, environmental and economic systems, and each has an impact on the other.
For this reason, Danone and XPRIZE have joined forces to design and develop a circular economy-based framework for a series of challenges aimed at meeting the UN SDGs. Companies such as Danone have looked at AI as a tool to build sustainable food resilience and invite local, diverse participation.
"The food system is so diverse that addressing one challenge will not solve the future. That is why we will work together on developing a suite of challenges and invite entrepreneurs to participate and work on creating solutions for the future," said Nigyar Makhmudova, Executive Vice President of Growth and Innovation at Danone. AI is seen as an "enabler" for good throughout this process, she said.
"The circular economy for food values framework will stand on three core principles: the food in this 'future of food' system should be inclusive and equitable; nourishing and desirable; and finally, regenerative and resilient by design," Makhmudova added.
How to participate: Continue the discussion and join the Circular Food Economy community to learn more about contribute to this effort.
Call to Action 3: Join the AI & Data for Good Global Alliance
Collaboration has been a common theme across the gender, food and pandemic tracks during the Breakthrough Days. Individual projects led by different teams benefit from building on the knowledge and work of one another.
This experience is why XPRIZE has decided to launch an AI and Data for Good Global Alliance in collaboration with industry leaders like Microsoft and Cognizant. It will be a "coalition of the most forward-thinking organizations and innovators focusing on enabling and cultivating actionable and tangible solutions to some of the world's most immediate challenges through the use of data sciences, machine learning, artificial intelligence for humanity and our sustainable future," Tauhert said.
How to participate: The Alliance is due to launch later this year. Members of the global AI and data community can express their interest by signing up here.
Call to Action 4: Help create a new ethical gender database
When artificial intelligence draws on male-dominated datasets, there is a real risk that it will amplify bias to create more inequalities instead of facilitating a better world for everyone.
Closing this gender data gap by collecting disaggregated and balanced data is essential for AI to provide thriving outcomes. To bridge that divide, XPRIZE has announced a new goal to create GenderNet - a comprehensive, diverse, ethically sourced and unbiased dataset that is made available for everyone to use in developing their algorithms.
The idea has been inspired by the now famous ImageNet, a large public dataset of images used to train algorithms to recognize objects. "When Fei-Fei Li started this initiative, it was a very unique approach to solving the problem of data. But fundamentally she believed that if we do not have a very good set of clean data that is properly tagged and usable, then AI advancement will not happen in time and it will not represent the real world. We are trying to do the same thing to solve the gender data gap issue," said Anousheh Ansari, CEO of the XPRIZE Foundation.
This endeavour will not just help eliminate gender biases in data, but will also detect and identify existing biases, Ansari said. "Together we can build something that is comprehensive, secure, safe and ethical. And it can be made available to the world so we can solve this problem once and for all," she added.
How to participate: Those who would like to become part of this initiative and help shape it are welcome to email their thoughts to firstname.lastname@example.org.
Image credit: fauxels via Pexels