Continuing the AI for Good Conversation: Takeaways from the 2018 AI for Good Global Summit
In May, the second AI for Good Global Summit brought leading experts together in Geneva, Switzerland, to demonstrate the multidimensionality of responsible AI applications for advancing sustainable development and humanitarian practice.
The Summit focused on spurring collective action on AI under four tracks: AI and satellite, AI and health, AI and smart cities & communities, and trust in AI. Under each track, breakout teams proposed AI strategies and supporting projects, guided by an expert audience representing government, industry, academia and civil society.
UN Global Pulse attended the Summit, sharing expertise from AI projects built with partners, and engaging in conversations to support ongoing innovation efforts and partnerships. Amidst the insightful presentations and lively discussions on bringing AI impact to scale, there emerged a number of key takeaways.
AI should augment human analysis, not replace it
In humanitarian scenarios -- like remote sensing analysis in the aftermath of a disaster, or satellite analysis of refugee settlements -- AI technologies can be used to identify objects in a fraction of the time human analysis requires. However, automation may fail to both detect certain nuances and to interpret results based on context and changing circumstances, which is critical to supporting operations on the ground.
Participants present at the Summit underscored that while automated learning and classification algorithms are vital to accelerating achievement of the 2030 Agenda, they should only serve as enablers to augment human analysis.
“For now, the role of analysts will remain to point the AI to the right questions to be analysed and to decide how to apply that analysis to problems in the real world,” said Dr. John Quinn, AI Adviser at Pulse Lab Kampala.
John leads the team that is currently working with UNOSAT (the United Nations Institute for Training and Research Operational Satellite Applications Programme) to develop automated algorithms that can assist human analysts to more efficiently and effectively count and analyse structures within refugee and migrant settlements.
Satellite imaging is benefiting from applications of AI and machine learning to train computers to interpret images
Speaking during the AI and satellite track, John added that “when dealing with conflict in humanitarian scenarios, precision in satellite image analysis is key. Applying a semi-automated approach that combines the efficiency of automation with the precision of human analysis helps us deliver results.”
AI can help bridge the digital divide and create an inclusive society
Greater democratization of AI can ensure that the benefits derived from the use of new technologies can reach everyone, especially people in developing countries. The rapidly-evolving capabilities of AI are already driving transformation across many domains. Mobility data from mobile phone networks can predict the spread of infectious diseases. Roofing materials visible from space can serve as a proxy for poverty, while postal records help estimate trade flows.
In remote rural areas where coverage and access to other forms of connectivity is limited, data mining techniques applied to what people say during radio talk shows can reveal opinions related to public healthcare, the state of roads and schools, or people’s trust in government.
The Radio Content Anaysis Tool works by converting public discussions that take place on radio into text
Today, most of the AI industry is increasingly being mined by businesses to predict consumer behaviour, track emerging trends, and monitor operations to improve sales and profit margins. Effective ecosystems for innovation and collaboration are needed to scale AI solutions for public good which are both gender and geographically diverse.
As a first step towards creating a mechanism for wider implementation, participants at the Summit committed to working together to promote cooperation, reduce duplication, and operationalize at scale some of the projects presented at the Summit. It was suggested that to strengthen internal institutional knowledge and expertise, organizations should create multidisciplinary data innovation teams of international and local experts to ensure projects are designed with a user-centric approach.
Participants also highlighted various initiatives and programmes that are helping inform, educate and empower citizens, especially the young, to include them in co-creating the AI solutions that they would ultimately benefit from.
The responsible use of AI will determine its impact
Throughout the Summit participants consistently reiterated the need to protect the public from harms arising from misuse of data and algorithms. They argued that in order to reap the benefits of AI, data-driven innovations must ensure transparency and accountability at every stage of project development, from collecting the data, to designing the algorithms, to implementing the solutions.
Three key dimensions of building trust in AI were identified during the Summit: AI stakeholders’ trust in AI developers; trust across national, cultural and organizational boundaries; and trust in AI systems themselves. These proposals will be supported by an incubator for multidisciplinary collaboration, namely the TrustFactory.ai.
Discussions also centred on how to include ethics as an additional tool to protect fundamental human rights. Efforts such as the EU’s General Data Protection Regulation and the United Nations Guidance Note on Big Data for SDGs: Data Privacy, Data Protection and Ethics, were mentioned to show progress towards mitigating the potential risks and harms of data use.
UN Global Pulse Director presenting the UNDG Guidance note on Data Privacy, Ethics and Protection
Examples were provided of companies who are partnering with governments, researchers and communities to advance data for social good initiatives and develop applications which focus on inclusion, safety and privacy.
Capturing the spirit of the conversations at the Summit, UN Global Pulse Director Robert Kirkpatrick added that as global efforts to develop new frameworks around the responsible use of emerging technologies begin to take shape, it is imperative that they address not only the human rights implications of “misuse”, but also those of “missed use.”