Proceedings of a Workshop

DATA-INFORMED SOCIETIES ACHIEVING SUSTAINABILITY: TASKS FOR THE GLOBAL SCIENTIFIC, ENGINEERING, AND MEDICAL COMMUNITIES

Proceedings of a Workshop—in Brief

The 2030 Agenda for Sustainable Development, adopted in 2015 by all United Nations (UN) Member States, offers a “shared blueprint for peace and prosperity for people and the planet, now and into the future.”¹ The Agenda outlines 17 Sustainable Development Goals (SDGs), which address a range of global challenges, including poverty, inequality, climate change, and environmental degradation, among others. Advances in technology and the proliferation of data are providing new opportunities for monitoring and tracking the progress of the SDGs. Yet, with these advances come significant challenges, such as a lack infrastructure, knowledge, and capacity to support big data.

To further examine how the global scientific, engineering, and medical communities can better facilitate the effective use of data to advance sustainability in the context of the SDGs, the National Academies of Sciences, Engineering, and Medicine’s Board on Research Data and Information (BRDI) and the Science and Technology for Sustainability (STS) Program convened a virtual public workshop on September 9–10, 2021. The workshop examined current efforts and initiatives to harness data and data-driven services to advance sustainability around the world. Workshop discussions also explored crosscutting issues, including strengthening the engagement of scientific, engineering, and medical communities on data-related issues, addressing disparities in the ability of societies to utilize data, and lessons learned from global experience with the COVID-19 pandemic.² Issues discussed during the workshop but not covered in depth that are important for future work include: the limitations to data availability especially in developing countries; the importance of ensuring data quality; the risk of creating datasets in a way that produce misleading results; social implications of data use such as increasing inequities among populations; and the recognition of the real-life challenges of implementation and delivery.

WELCOME AND WORKSHOP GOALS

Thomas Arrison, National Academies, welcomed participants to the workshop and discussed current efforts to improve the collection and utilization of data relevant to the SDGs, particularly around tracking of progress towards the goals to inform decision-making. Research on complex, multidisciplinary, and transdisciplinary issues central to achieving the SDGs increasingly utilizes large datasets and machine learning. This workshop provided an opportunity for a global group of experts to discuss challenges and opportunities for the scientific, engineering, and medical communities to strengthen utilization of data necessary to achieve the SDGs.

Nebojsa Nakicenovic, International Institute for Applied Systems Analysis, and Workshop Chair, discussed how current complex global challenges, such as the COVID-19 pandemic and climate change, are informed and changed by data. The goal of the workshop was to explore how the global scientific community can facilitate the effective use of data to advance sustainability in the context of the SDGs. Specifically, the workshop would encourage the strengthening of the engagement of the scientific community in efforts to shape the post-2030 UN Agenda on data-related issues; “the world will not stop in 2030.” The digital age is crucial for the achievement of sustainability for all, Nakicenovic said.

__________________

¹ See https://sdgs.un.org/2030agenda.

² For an agenda, selected speaker presentations, and a video of the workshop, see https://www.nationalacademies.org/event/09-09-2021/data-informed-societies-achieving-sustainability-tasks-for-the-global-scientific-engineering-and-medical-communities-a-virtual-workshop.

SETTING THE STAGE: DEVELOPING DATA-INFORMED SOCIETIES TO ACHIEVE SUSTAINABILITY

Philip Bourne, University of Virginia, provided framing remarks around how data can and is advancing sustainability. Data has been referred to as “the new oil” or “new soil,” given its importance to society. Data science, or as Bourne described it, the intersection between math, statistics, and computer science, touches nearly every field and domain (Figure 1).

Bourne discussed a University of Virginia project related to UN SDG 10, focused on “reduc[ing] inequality within and among countries,” where students and researchers are examining parish records to understand social networks of Native Americans in the Northeast. The goal is to examine issues around migration, including how we can manage and control migration in a positive way. The project offers an example of the usefulness of big data in examining these kinds of issues; no field will be studied in quite the way it has been in the past given the advent of large amounts of data, Bourne said. However, he cautioned, “big data is not a magic bullet.” While it may broaden opportunities, its collection and use need to be carefully considered and approached through an interdisciplinary lens. Artificial intelligence (AI) and machine learning offer examples of the potential of big data, but should also be approached with caution, given what is known about the potential for bias and disparities in underlying data that may support this technology. Bourne noted the importance of conducting a Strengths, Weaknesses, Opportunities, and Threats (SWOT) analysis for assessing use of big data.

Citizen science can and is playing an important role in big data, said Bourne. With a connection to the internet and a device, one can access a large amount of open and public data and make contributions. However, there is a need to build trust among the public about these approaches and big data in general. Translating data into action is also a challenge, Bourne noted. We often think of science, technology, engineering, and medicine (STEM) as the nexus of development from which we can develop policy and law to drive change. However, data are messy and can be misinterpreted, thus, making this translation difficult. Lack of training and the proprietary nature of data are also barriers to using data to drive decision-making. Sustaining an open data ecosystem is also going to be a significant issue.

While optimistic about the implications of rapid expansion of data, Claire Melamed, Global Partnership for Sustainable Development Data, recognized the significant barriers that exist in terms of the use of big data. There are important inequities that must be acknowledged around its collection and use; these also affect power relationships. Thus, the potential benefits and likely harms of big data are going to be unevenly distributed. Quoting William Gibson’s “The future is already here, it’s just not very evenly distributed,” Melamed added that to address these challenges, we need to consider what future we want.

Through data and technologies, we have a chance to reduce inequality, Melamed said. However, we need to understand drivers of inequality, both individually and collectively, and take the steps to address them. One way to do

this is to bring together different disciplines—bridging data science and social science in different dimensions. Without a common goal, either at the national or global level, new technologies will enter the system and continue to reinforce old inequities. There is a need to keep people at the center of the analysis and ask questions about what it may mean to be left out of the data and how does this exclusion drive inequality? Considering who develops the technology and creates algorithms is also important as those involved may not be aware of inherent bias that may be introduced in underlying data. There is a need for a system that examines how politics and markets impact data and drive inequality, Melamed noted. Ultimately, data is knowledge—and knowledge is power. She added, “we need to consider how we can give as much of this power as possible to people and not take it away.”

BIG DATA, DIGITIZATION, AND COVID-19: LESSONS LEARNED AND WAYS FORWARD

During the first session, panelists discussed issues related to big data, digitization, and the COVID-19 pandemic, noting lessons learned and efforts around promoting equity in sustainability. Barend Mons, Leiden University Medical Center, moderated a panel addressing lessons learned related to big data and the use of digital technologies during the pandemic. He began stating that “knowledge is like love, it multiples when shared. When I share love with someone, I don’t lose it.” Most of the challenges we face around data and open science are concerns about sharing, he said.

The COVID-19 pandemic highlighted the critical need to examine the proliferation of disinformation and misinformation online, said Sheldon Himelfarb, PeaceTech Lab. He cited an example from the Center for Countering Digital Hate that found that 65 percent of all anti-vaccine content shared or posted on Facebook and Twitter between February and March of 2021 can be attributed to just 12 individuals.³ This contributed significantly to vaccine hesitancy and ultimately, needless deaths. This spread of misinformation is becoming an existential threat to the planet, comparable to other threats, such as climate change and future pandemics.

There are several existing proposals to try to address the spread of online misinformation, Himelfarb said. These have focused on regulating social media platforms, requiring them to act faster and more effectively to pull down content, or punishing those who spread misinformation online. There are also proposals that focus on educating the consumer or requiring media literacy curriculum in schools. However, this is a global problem that disregards borders and requires a global response. To this end, Himelfarb noted that he and others have called for the creation of an Intergovernmental Panel on the Information Environment (IPIE) to address the misinformation crisis. Modeled on the Intergovernmental Panel on Climate Change (IPCC), an IPIE would analyze the global information environment and provide science-based recommendations for achieving vital information integrity.⁴

Carly Kind, Ada Lovelace Institute, discussed her focus on the ethics and social implications of data and new technologies, particularly AI. Access to data in and of itself is not the key challenge as we work to address the SDGs. In fact, the key issues include lack of data infrastructure, along with limited data literacy. Also of concern was the lack of standardized approaches for data protection, particularly around transferring data from one jurisdiction to another in the context of different data protection rules. In the absence of a set of an international standard, easy and rapid sharing of data has been impeded, Kind noted.

The lack of public trust in the government was also highlighted during the pandemic. Without this trust, we face barriers in adopting data driven approaches to the SDGs, even in the face of data access issues, Kind added. Rebuilding public trust in the use of data for social problems is needed, said Kind, and public involvement is vital. Governments have failed to articulate a vision for the future in which data plays a central role; data should be treated as a public good that is used for the public interest.

Francisca Onaolapo Oladipo, Federal University Lokoja, and VODAN (Virus Outbreak Data Network) Africa, described a comprehensive open data initiative in Africa during the pandemic. A collaborative of African countries, including universities, ministries of health, and health facilities in Uganda, Ethiopia, Somalia, Tanzania, Kenya, Nigeria, Liberia, Burkina Faso, Zimbabwe, and Tunisia, worked to effectively support data collection around COVID-19 infections.⁵ Ensuring that data are findable, accessible, interoperable, and reusable (FAIR) is critical to help inform decision making.

Oladipo noted that there are several social barriers related to the collection and use of data, particularly around collecting sensitive health data, including a lack of health management information systems. There is a need to develop an initial data infrastructure and training around technologies to allow for data sharing. Foundations and the private sector are key partners; for example, Google and other private sector entities provided research funding

__________________

³ See https://www.counterhate.com/disinformationdozen.

⁴ See https://www.peacetechlab.org/news/2021/5/17/peacetech-lab-leads-tech-and-human-rights-experts-in-call-for-bold-action-to-counter-misinformation-during-nobel-prize-summit.

⁵ van Reisen et al. 2021. Design of a FAIR digital data health infrastructure in Africa for COVID-19 reporting and research. Advanced Genetics 2(2):e10050. https://doi.org/10.1002/ggn2.10050.

to develop and support the data infrastructure for the project. The global challenges outlined in the SDGs require the full participation of all to develop and share data. The COVID-19 pandemic has demonstrated the importance of data sharing for the good of public health and a unifying framework on privacy and data protection could support a broader data sharing effort, noted Oladipo.

Wayne Koff, Human Vaccines Project, discussed the Project’s role as a nonprofit research consortium with laboratories all over the world dedicated to vaccine research. During the pandemic, the speed with which the COVID-19 vaccine was developed was widely held as a success. Yet, despite this success, we continue to see millions of cases and deaths due to the virus.

One key initiative underway is an effort to develop an AI model of the human immune system to assist with vaccine research. AI and machine learning in the informatics space have provided opportunities to understand the human immune system in new ways that even a decade ago people would have thought were unfathomable. However, the challenge we face today is that biologists and the AI and machine learning researchers all speak different languages, Koff said. A vision or ambitious goal is needed to expand interdisciplinary efforts for big data. While an AI model of the human immune system may be a possibility, a long-term goal and vision is needed to move society in this direction.

Discussion

Panelists discussed the need for social scientists, regulators, and researchers to work together to address issues around data integrity and privacy. Additionally, panelists discussed the current lack of a data infrastructure, for example, to support data exchange. Funding is necessary to support this infrastructure as well as training for researchers, including around communications skills. Discussing the lack of trust in public institutions as it relates to the use of data, Kind described the important role of data stewardship, an oversight role that is responsible for the governance and responsible use of data. There is a need to insert data stewardship as a practice at various layers throughout society, starting with scientists and researchers, she said. “Better policing of bad actors in order to create more trust in good actors is also really important,” said Kind. To this end, participants also reinforced the idea of establishing a commission or body that set standards for data to prevent the spread of misinformation.

BIG DATA AND DIGITALIZATION FOR PROMOTING EQUITY AND EQUALITY IN SUSTAINABILITY EFFORTS

Luis Bettencourt, University of Chicago, moderated a panel addressing the role of big data and digitalization in promoting equity and equality in sustainability. While COVID-19 accelerated technological development, it also exposed inequalities. The spread of information and communication technologies created immense value; however, access is uneven. Bettencourt noted that there is a need to utilize data and technology to support sustainable and equitable change, inquiring, “How can these technologies and data act as both equalizers as well as engines of development?”

The use of digital technologies in science is pervasive in all domains of science, said Heide Hackmann, International Science Council (ISC). Hackmann discussed several threads that cut across the sustainability challenges we are facing, including pathways to global sustainability as the biggest and most urgent challenge. We have entered a new era of open science, built on the capacity of digital infrastructures, Hackmann noted. The ISC is engaged in several related initiatives around open data. The Coalition for Digital Environmental Sustainability, or CODES, is a global community of practices that aims to develop and implement an acceleration plan for action at the nexus of environmental sustainability and digitalization.⁶ The Council is also active in articulating and advocating for principles of open science.

Hackmann also discussed a program launched in 2019 that positions scientists and science systems in low and middle-income countries in the Global South at the cutting edge of data-intensive open science.⁷ The African Open Science Platform hosted by South Africa’s National Research Foundation is serving as a model for the development of analogous platforms in other regions. By harnessing the tools of the digital revolution, we are fueling scientific discovery, Hackmann said. However, there is a need to raise awareness and, if possible, safeguard against potentials of digital technologies and processes that deepen inequalities, exclusion, and discrimination. There is a need to address inequalities in global science by developing individual, institutional, and political capacities for data-intensive open science. Challenging and changing the governance of science through an emphasis on policies related to openness, inclusivity, and diversity are needed, Hackmann stated.

Sheela Patel, Society for the Promotion of Area Resource Centers, discussed her organization’s work to address poverty in India, particularly the use of big data initiatives. Patel noted that people living in poverty are often not

__________________

⁶ See https://www.sparkblue.org/CODES.

⁷ See https://council.science/actionplan/open-science.

engaged in data collection or are ignored in the process, thus, facing additional barriers in terms of having to produce evidence to justify their demands. The exclusion of the poor in health systems data is evidence. Food supply systems in India are also exclusionary; these systems provide support only those who can produce identity cards. While national data are being collected and analyzed to assess evidence on the SDGs, Patel noted that these efforts are not capturing information about the impact on those living in poverty. When the data are collected, there have been instances when the government will not accept it, highlighting the role of politics. Additionally, those communities who participate in data collection do not often learn about the results or benefit; open data is not equally available to everyone. “Real data is messy, cluttered, and full of problems,” posing long-term challenges for those involved in addressing the SDGs, said Patel. There is a need for a long-term commitment to using data to transform the lives of those in poverty.

The pandemic has been devastating in terms of its impact on global poverty. Joshua Blumenstock, University of California, Berkeley, discussed an example of one country’s efforts to use data to develop a targeted poverty relief program. The government of Togo developed a data-informed program, called Novissi Program, to prioritize aid for poorest people in the country. Through a new platform, the government was able to distribute aid through mobile phones and mobile money. Initially, the government did not have the social registry or data to support this type of targeted antipoverty program, particularly as its last census was in 2011. Blumenstock worked with the government to develop the data infrastructure to identify and prioritize need through high-resolution satellite imagery with machine learning technology (see Figure 2).⁸ The Togolese government also sought to distribute a $10 million private gift to 150,000 of the poorest cantons. Blumenstock, in collaboration with the nongovernmental organization that had provided the financial support, used mobile phone metadata and machine learning to help identify recipients. With crises, there is no time to conduct nationwide surveys to determine need. This example shows the power of satellites and cell phones in rapidly generating data.

The work in Togo also illustrates the potential for the applications of big data in sustainable development. Despite the challenges of reaching people without mobile phones, providing data access and data privacy, and safeguarding against future (mis)use, the use of data should complement, not replace, existing methods in targeting humanitarian aid, Blumenstock discussed. Community-based methods also provide additional tools. Each approach has strengths and weaknesses, but the goal is ultimately to consider how to integrate these into a layered system for social protection, which also includes mechanisms for appeals and recourse and informed consent, Blumenstock said.

Richard Heeks, University of Manchester, discussed his work to analyze the impact of digital technologies on international development efforts,⁹ beginning with a discussion of innovations to address data roadblocks. He described a project supported by the German Agency for International Cooperation (GIZ) and the United Nations Development Programme (UNDP) Accelerator Labs Network that is using big data to identify development outperformers; for example, farmers who are significantly more productive than their peers at conserving their land. New datasets can be used to understand factors that may contribute to this outperformance. However, Heeks noted that converting this data into information that guide decision making can be slow and costly due to formatting issues and insufficient quality. There may also be barriers in encouraging development agencies to use new data. To address these issues, efforts are needed to move more data into the public domain, create tools to support the cleanup of datasets, and to train scientists and decision makers around big data, particularly in developing countries.

__________________

⁸ Chi, G., H. Fang, S. Chatterjee, and J. E. Blumenstock. 2021. Micro-Estimates of Wealth for all Low- and Middle-Income Countries. https://arxiv.org/abs/2104.07761.

⁹ Heeks, R. 2021. From Digital Divide to Digital Justice in the Global South: Conceptualising Adverse Digital Incorporation. Digital Development Working Paper 90. Available at https://www.gdi.manchester.ac.uk/research/publications/di/di-wp90.

Regarding education and training, the University of Manchester has been working as part of a university partnership to build data science education in Columbia.¹⁰ Through this initiative, Heeks noted that training needs greater integration of real-world data experiences. A data fellowship model that places students into organizations could better support required data capacity building.

Heeks also discussed the idea of data justice as an important goal moving forward. Data injustices can arise from exclusion, including the digital divide, and particularly affect the Global South. While there is growing interest in use of smartphones and social media data to guide development decisions, low-income populations, women, ethnic minorities, and older people are less often users of cell phones or social media and thus are among those who are significantly underrepresented in derived datasets. Heeks added that there is also a need to understand how inequality derives from inclusion in digital data systems, known as adverse digital incorporation. To combat this, we need to work towards creating advantageous digital incorporation: digital interventions that specifically seek to reduce existing contextual inequalities in order to deliver data justice (see Figure 3).

Heeks noted that existing contextual inequalities of money, skills, and technology access, along with control over data rights create significant challenges for delivering data justice for sustainable development. To address this, data rights that are enshrined into law are needed. In particular, marginalized groups in the Global South must be fairly represented in datasets, have the right to access and correct data about themselves, and share right of ownership over that data, or at least over the benefits accruing from it.

Discussion

Panelists discussed how to better acquire the perspective of individuals and communities to identify solutions to sustainable development challenges. Patel noted that the first step is to gain a clear understanding about the roles of all involved, including working to develop relationships with the community. Regarding the role of scientific organizations, Hackmann discussed a need for global transdisciplinary research and deeper engagement toward equitable and sustainable development. Panelists also discussed the importance of collaboration, knowledge creation, incentives among researchers and data providers, and capacity building strategies through an interdisciplinary approach, in addition to education and training for the next generation.

BIG DATA, DIGITIZATION, AND THE SUSTAINABLE DEVELOPMENT GOALS

Panelists during the second session discussed broad issues related to digitalization, big data, and the SDGs, including the role of big data in our digital future.

__________________

¹⁰ Jones, P., J. Carter, J. Renken, and M. A. Tobón. 2021. Strengthening the Skills Pipeline for Statistical Capacity Development to Meet the Demands of Sustainable Development: Implementing a Data Fellowship Model in Colombia. Digital Development Working Paper 89. Available at https://www.gdi.manchester.ac.uk/research/publications/di/di-wp89.

¹¹ Based on https://www.paperpinecone.com/blog/teaching-difference-between-equality-equity-and-justice-preschool, adapted from https://designintech.report/2019/03/11/%F0%9F%93%B1design-in-tech-report-2019-section-6-addressing-imbalance.

Digitalization, Science, Technology, and Innovation (STI), and the SDGs: What Should Be Measured and How?

Aleksandra Berditchevskaia, Centre for Collective Intelligence Design, Nesta, moderated a panel focused on digitalization, STI, and the SDGs, noting that while the SDGs provide an overarching framework, there are significant challenges around how they are defined and measured. Before it becomes possible to measure progress, we must face the complexity of establishing baselines, and understanding how best to evaluate what is meant by success, she stated. There is currently a lack of congruence between how progress and action are measured at different scales, meaning efforts are often piecemeal without a pathway for integration into the institutional infrastructures around the SDGs. Making sense of these inconsistencies and mapping connections, including around governance and decision making, is a daunting task. It will require methodological innovations, including those discussed below, that draw on collective intelligence to bring together diverse groups of people, novel sources of data, and technology such as AI to examine different scales of sustainability, from local to global.

Sarah West, Stockholm Environment Institute, discussed the role of citizen science in supporting the SDGs. Citizen science, defined as research in which volunteers work with scientists to answer real-world questions, can take place at different scales, and may include volunteers at all stages of the scientific process from developing a hypothesis and collecting data to having useful information (Figure 4). The approach has become popular for many reasons, including a push from funders and governments for scientists to engage with the public more fully. Potential benefits of this approach include the ability to collect data over large geographic areas, have educational outcomes alongside scientific ones, and that projects with a diverse range of participants can be more creative in both project design and in suggesting practical solutions, among others.

Regarding the role of citizen science in the SDGs, West cited a 2017 discussion brief written with her colleague Rachel Pateman on defining, monitoring and implementing the SDGs through citizen science.¹² She provided an example of the Tupumue project¹³ where citizen science is contributing to understanding around air pollution and lung health in Nairobi, where they are using air sensors to collect air pollution data along with spirometry testing, questionnaires and children’s stories. The project was developed in coordination with the community. West noted that citizen science has raised awareness about sustainable development more broadly, and the SDGs.

While an important research approach, West added that citizen science has its limitations. It struggles with a lack of diversity in its participants, with mostly white and educated participants, not representative of the population as a whole. As an approach, it generally relies on those who already have access to technologies, which may be contributing to a less diverse set of participants. Citizen science requires robust procedures to ensure the quality of the data. There is also the larger challenge about how to incorporate these data into national data sets.

Joy Bonaguro, State of California, discussed the state’s efforts related to collecting and sharing data, including its statewide data goals, and work to streamline data access, improve data management and governance, and spur data use. “We can’t measure what we can’t access,” a key challenge given that data are collected by various institutions and for a variety of purposes, however, these efforts are often siloed. “If you are only able to see part of the problem, you can’t come up with a holistic approach” to address it. The state has developed several initiatives to help break these siloes, including for example, linking early childhood and workforce data as well as developing a unified homelessness data system.

An additional challenge is the impact of existing rules around data privacy, both federally and locally, that make the exchange of data complicated. “Even if we are able to gain access to data, we can struggle to use it.” To

__________________

¹² West, S., and R. Pateman. 2017. How could citizen science support the Sustainable Development Goals? Policy brief. Stockholm Environment Institute. Available at https://www.sei.org/publications/citizen-science-sustainable-development-goals.

¹³ See https://www.sei.org/projects-and-tools/projects/tupumue.

address this, Bonaguro noted, the state of California is working to articulate data standards and build a data architecture. The need for capacity building to support these efforts in government is significant. She described a “solution typology” that is designed to develop capacity using advanced analytics in program management in local government: (1) find the needle in the haystack, (2) prioritize your backlog, (3) flag “stuff” early, (4) AB test something, and (5) optimize your resources. Through this solution typology, Bonaguro has helped program managers to incorporate advanced analytics in their work, including through efforts to process their backlogs and are prioritizing what needs to be addressed first. To advance some of these issues, Bonaguro noted that the U.S. Department of Energy’s national laboratory system¹⁴ could serve as a model for the social sciences or government sciences. At the state level, each entity is currently developing their own data infrastructure and building capacity, with no structured or ongoing partnership. A national laboratory system could better support government agencies in managing and utilizing their data, she noted.

Miguel Luengo-Oroz, United Nations Global Pulse, described the UN’s expansive use of nontraditional tools to measure issues such as mobility and perceptions, including in humanitarian settings as refugees camps. The UN can aggregate data to support leaders to make well-informed policy decisions or determine how policy responses are working (See Figure 5 for the evolution of digital data used for decision making between 2013 and 2019). For example, during the COVID-19 pandemic, aggregated and anonymized data from cellphone towers has been used to determine how effective lockdown policies were in various locations in the European Union and other countries. Also, advanced speech technology has been used to assess how misinformation is spread, for example, about vaccines. However, this technology is not available to capture the nearly 7,000 languages spoken today—digital technologies have been developed worldwide for about 100 languages. The models are not coded to assess these languages, and thus are inherently biased and risk leaving some voices behind and amplify inequalities.

With this rapid data innovation comes additional benefits, but also risks and potential harms, as machine learning techniques have allowed for additional spread of misinformation and disinformation, Luengo-Oroz said. Beyond amplifying the spread, deep learning techniques allow creating fake visual (or “deep fakes”) and textual digital content, which will not be distinguishable from real data, generating new threads for societies.¹⁵ For example, UN Global Pulse researchers were recently able to train AI to fake UN speeches in just 13 hours.¹⁶

Luengo-Oroz noted that the UN’s Secretary General launched Our Common Agenda in September 2021¹⁷ with specific proposals designed to accelerate the achievement of the SDGs, including a commitment to “improve digital cooperation” through a Global Digital Compact. Digital cooperation efforts aim to connect all people to the Internet, including all schools, avoid internet fragmentation, protect data, apply human rights online, introduce accountability criteria for discrimination and misleading content, promote regulation of AI, and create digital public goods.

Alexandre Caldas, United Nations Environment Programme (UNEP), offered several lessons learned around integrating data and knowledge into his own organization’s efforts, including the need for leadership, sustainable

__________________

¹⁴ See https://www.energy.gov/science/science-innovation/office-science-national-laboratories.

¹⁵ For more information on deep learning and deep fakes, see Nguyen, T. T., Q. V. H. Nguyen, C. M. Nguyen, D. Nguyen, D. T. Nguyen, and S. Nahavandi. Deep Learning for Deepfakes Creation and Detection: A Survey. https://arxiv.org/abs/1909.11573.

¹⁶ Bullock, J., and M. Luengo-Oroz. 2019. Automated Speech Generation from UN General Assembly Statements: Mapping Risks in AI Generated Texts. https://arxiv.org/abs/1906.01946.

¹⁷ UN. 2021. Our Common Agenda—Report of the Secretary-General. Available at https://www.un.org/en/content/common-agenda-report.

funding, a centralized approach to the collection and use of data, and strong partnerships. The correct identification of target audiences and stakeholders is also important, including attention to the needs of policy makers, the scientific community, international organizations, business and innovators, and citizens and civil society. Integration, impact, and partnerships are all necessary to building and supporting data capacity.

To enable innovation in data capabilities, Caldas noted that the UN is working on several key and related cross-cutting agendas, including a data strategy that encourages maximizing the value of data as a strategic asset; an innovation agenda to allow for the discovery and implementation of ideas to create value; an examination of the use of digital technology to change business models; a behavioral science agenda to allow for insights into policies and programs; and strategic foresight capability, for use in planning for future actions.¹⁸

As part of this work, the UNEP has developed a roadmap for implementation, through a dashboard of services, focused on seven channels with core data structured around these channels (see Figure 6).¹⁹ The channels include (1) geospatial, technology and knowledge platforms; (2) multilateral environmental agreements, SDGs, and environment statistics; (3) scientific information on the environment; (4) global environment monitoring systems; (5) global environmental outlook; (6) citizen science; and (7) strategic foresight. As part of data integration efforts, UNEP is currently engaged in over 15 demonstration projects, including related to ozone, oceans, climate change, and others, crossing divisions of the UN. The agency is examining how data elements can be combined to provide critical information to achieve the SDGs, as well as to inform decisions and action on the ground. He noted that while access to data is priority, a secondary priority is the intelligent analysis of data; “we need data analysts more than we need data.”

Caldas also discussed the World Environment Situation Room (WESR),²⁰ which implements the Big Data Initiative, a global project that includes geo-referenced, remote-sensing, and earth observation information integrated with statistics and data on the environmental dimension of sustainable development. Through this effort, global monitoring and other tools are examining pollution, and other issues such as climate change and changes to biodiversity. The platform allows for interaction between various data tools and users, including the ability to create story maps.²¹

Discussion

Panelists discussed the tension between opportunities and challenges in working to engage government institutions to pay attention to data sources. As West noted, what is lacking is communication between those who are designing and commissioning studies and those conducting data reporting. In citizen science, there has not been enough thought given to how this research can fit in with the SDGs. Government statistical offices need to talk with citizen science

__________________

¹⁸ In addition to Our Common Agenda—Report of the Secretary-General, see UN. 2020. Data Strategy of the Secretary-General for Action by Everyone, Everywhere with Insight, Impact and Integrity: 2020-2022. Available at https://www.un.org/en/content/datastrategy/images/pdf/UN_SG_Data-Strategy.pdf.

¹⁹ UNEP. 2019. Digital Transformation towards a Global Environmental Data Strategy: People, Places and Planet. Available at https://wedocs.unep.org/bitstream/handle/20.500.11822/29769/DigitalTransformation_GlobalDataStrategy_ReportCPR_10Dec2019.pdf.

²⁰ See https://wesr.unep.org.

²¹ Caldas demonstrated several tools available on the WESR website, including world population real-time exposure to air pollution (https://www.iqair.com/unep), global monitoring (https://wesr.unep.org/topic/globalmonitoring), and the evolution of Arctic sea ice extent (https://graphs.unepgrid.ch/graph_arctic_sea_ice.php).

practitioners; this has the potential to lead to action.

Negotiating partnerships and allowing actors to engage and communicate is important, noted Bonaguro, starting with problems and moving toward solutions. Building trust in these partnerships is also critical, Caldas said, along with attention to a holistic approach to developing and building the data infrastructure. Ultimately, we need organizations to talk to each other; the domains and siloes need to be bridged. In addition, one panelist questioned whether more data are needed or if there is a need to handle the data we have more effectively.

ROLE OF BIG DATA AND COLLABORATION IN OUR DIGITAL FUTURE

Moderating a panel addressing the role of big data and our digital future, Jeanne Holm, City of Los Angeles, noted that there is a need to consider the connection between the ethics of data, access to big data, the ability to use data for predictive analytics and what that means for our collective intelligence, particularly as we move toward the use of AI and machine learning. Picking up on many of the themes discussed in earlier parts of the workshop, Holm noted that partnerships are essential to this work. Despite rapid advancements in data collection, there are deep inequities have been exacerbated during the pandemic, including the digital divide and lack of digital access, described earlier. The idea of equity and big data and digital transformation all interplay. The question becomes, how we will transform and grasp the opportunity of digital transformation and relate it back to the SDGs?

Pedro Conceição, United Nations Development Programme (UNDP), began by discussion the aspiration of the UN development agenda, “to ensure that all human beings can fulfil their potential in dignity and equality and in a healthy environment.” This aspiration is critical to driving how we identify problems and solutions to those problems. Discussing the use of metrics to quantify progress on the SDGs, Conceição, noted that these can be useful in shifting policy but have limitations. In times of uncertainty, for example, during the pandemic, metrics can provide quantifiable information but often lag, limiting their applicability to policy decisions, which must be made in real time. Importantly, the UN is using a human development framework to address challenges, Conceição stated.²² Along with the Climate Impact Lab,²³ a global partnership around research quantifying the impacts of climate change, the UN is working to cross references vulnerability to climate change at the local level using data. Big data is providing the opportunity to examine behavior in real time, which has been transformational. Through satellite data combined with AI algorithms, it might be possible to assess poverty in real time.²⁴

Huadong Guo, Chinese Academy of Sciences (CAS), discussed his work in assessing big data used in the Earth sciences, including the CAS Big Earth Data Science Engineering Program. The Program is designed to develop a data-sharing platform to host big data and cloud services, to promote scientific discoveries in Big Earth Data, and to provide comprehensive decision-making support (as outlined in Figure 7).

__________________

²² UNDP. 2020. Human Development Report 2020: The next frontier: Human development and the Anthropocene. Available at http://www.hdr.undp.org/en/2020-report; and UNDP. 2020. 2020 Human Development Perspectives. COVID-19 and Human Development: Assessing the Crisis, Envisioning the Recovery. Available at http://hdr.undp.org/en/hdp-covid.

²³ See https://impactlab.org.

²⁴ See the Togo example presented by Blumenstock who developed the data infrastructure to identify and prioritize need through high-resolution satellite imagery with machine learning technology.

The International Research Center of Big Data for Sustainable Development Goals (CBAS), as described by Guo, was created to support innovation and application of data for the UN 2030 Agenda for Sustainable Development. The Center works toward a vision where data is open and accessible across border and disciplines. Key activities include developing the SDG data infrastructure and information products, launching a series of SDG satellites, providing new knowledge for SDG monitoring, establishing a think tank to promote the SDGs, and providing capacity development for SDGs in developing countries. The SDG Big Data Platform is designed to provide data resources and cloud services for SDG decision making. The Platform offers 10PB (petabyte) data covering multiple subjects and provides opportunities to characterize and profile scientific workflows and cloud services.

Jennifer Pahlka, Code for America, discussed the level of collaboration and the role of society, particularly volunteers, in addressing community need during the pandemic. In a striking indication of how unprepared the U.S. was for the public health emergency, civil service and volunteer organizations needed to step in to address critical gaps typically filled by the public sector, including in data collection and analysis. These efforts included supporting analyses to indicate where intensive care unit capacity was limited. In this case, more than 300 volunteers worked together with the public and private sector to collect these data, offering a true demonstration of the power of public-private partnerships. Pahlka described other examples of organizations who took the lead to address critical challenges during the pandemic, including a public-social sector partnership to help decision makers in criminal justice understand how to safely address over-incarceration during the pandemic, through the collaboration of engineers, data, and scientists.

While the pandemic has accelerated data driven decision making, there continues to be a need for capacity building, particularly in the public sector. Also, to address equity problems in the U.S., we need to ask different questions about our data and support the capacity of public agencies who are working to address these larger questions.

Himla Soodyall, Academy of Science of South Africa, said that in South Africa, the term SDGs are neither well known nor used to inform policy decisions. The SDGs are also poorly integrated into academic and scientific community. A key challenge is engaging the various stakeholders around addressing the SDGs, an effort which has been further delayed by COVID-19. Soodyall discussed the significant impact that the digital divide has had on education in South Africa. With COVID-19, the country is facing the triple challenge of unemployment, poverty, and inequality—one-third of the country’s population is unemployed, and poverty is dramatically higher.

Soodyall discussed South Africa’s Council for Scientific and Industrial Research, which maintains a national data observatory of resources integrating data at various scales.²⁵ The Academy is developing a code of conduct for research to guide issues related to the ethical use of data. Soodyall discussed several areas needed to make progress include addressing siloes of technology and information to layering of the information to support decision making. Our challenge is using evidence-based science in the service of society and in bringing ethics back to data.

Discussion

Panelists discussed the need to make data more accessible; the more we can make connectively from human suffering to opportunity is important. Access to digital technologies has shown a direct impact on issues such as education and health. As discussed in earlier parts of the workshop, panelists emphasized that the pandemic created a new sense of urgency for addressing the digital divide and highlighted the need to incorporate global cooperation in our projects, the need for the scientific community to come together to support big data efforts. Innovative data collaborations are also critical, as described by Soodyall, in Africa’s open science platform. Sustained resources are needed to support these projects moving forward. Government will never have all the capacity it needs to address significant challenges such as the pandemic. However, the crisis has shown the power of the community and volunteerism. Understanding user need is also important to making the greatest impact. Panelists also discussed the idea of transparency, both in terms of the ethical use of the data but around the data itself. An ethical code of conduct for researchers could help address some of these challenges. Also, data literacy or access can be an inhibitor for collecting, sharing, and using data, while Conceição discussed the importance of focusing on human capabilities.

PARTICIPANT REFLECTIONS ON DATA-INFORMED SOCIETIES ACHIEVING SUSTAINABILITY

Solomon Hsiang, University of California, Berkeley summarized the workshop discussions on how the global scientific community can better facilitate the effective use of data to advance sustainability in the context of the SDGs. One key theme from the discussion, as noted by Nakicenovic, is that data is both a great equalizer and great divide in society. Data in many ways is unique and different from other resources that are used around the world. Data, as Bourne states,

__________________

²⁵ See https://www.csir.co.za.

is the new oil and soil, however, as opposed to other resources, there are low barriers to entry. While big data are a global public good, exclusion and injustice through data is a new form of exploitation. Workshop presenters noted numerous threats in the face of big data, including its role in creating inequality. As discussed, there are questions around who is included and excluded from the data and how that may perpetuate inequities. Similarly, as Patel notes, vulnerable communities who may contribute to data collection may also not benefit from their efforts; the benefit from the data should return to those communities.

Another topic of important discussion during the workshop is the role of misinformation and disinformation, Hsiang said—a key problem that prevents us from solving other problems, such as, climate change and is contributing to inequality. Panelists discussed the need for a global response, specifically developing a commission or intergovernmental panel on global disinformation, such as the IPIE as suggested by Himelfarb. There are synergies across disciplines; breaking silos is important as is the need for integrated approaches and a holistic view of the future. We do not currently have an inclusive vision of the future we want. Central to this transformation is knowledge; as shared by Mons previously “knowledge is power” and like love, it is not lost through sharing. Hsiang reinforced this point, noting the need for cross disciplinary approaches to big data.

Hsiang summarized panelist comments around the use of nontraditional data for rapid intervention (for example, the novel data sources used during COVID-19) such as the use of satellites and cell phone data to alleviate poverty and advance progress on the SDGs, including the Togo example presented by Blumenstock. Consideration is also needed to ensure we can develop incentives among researchers and data providers to make sure that innovations and data are made accessible and integrated into official SDG monitoring infrastructures. However, we need to understand exactly what it is we are measuring. The ethics of and privacy challenges associated with collecting data during COVID-19 was highlighted and will continue to be a challenge moving forward, Hsiang noted. The importance of data quality was also discussed, including data generated in citizen science efforts that require robust procedures to ensure quality, as described by West. The pandemic demonstrated the value of data to decision makers; hopefully, this will change the way decisions are made in the future. A key question during the discussion also included whether more data are needed or if there is a need to handle the data we have more effectively, in addition to incentives and funding to create data infrastructure.

Hsiang added there may be a need to create an 18th SDG focused on designing, building, and maintaining data collection systems for the other 17 SDGs, as this is on its own is a major unprecedented global challenge. Addressing this will require ongoing infrastructure, interdisciplinary training, extensive public engagement, and significant resources. Developing a system where we can instantaneously track progress on the SDGs and identify what individuals can do to help advance these collective goals would be a vision for the future, Hsiang said. To conclude the workshop, planning committee chair Nakicenovic acknowledged a positive impact of the digital revolution and how the COVID-19 pandemic is changing our society, including an estimated 9.8 billion mobile phones, 2,200 satellites, and more than 25 billion other digital sensors that have allowed us to share information across the globe.²⁶ He noted the general optimism of participants that big data can catalyze the transformation over the SDGs, if not by 2030, then by 2050, and that it would be useful to develop roadmaps for a positive way forward to assess our digital common future.

__________________

²⁶ See https://www.wri.org/insights/insights-big-data-how-covid-19-changing-society.

DISCLAIMER: This Proceedings of a Workshop—in Brief was prepared by Thomas Arrison, Franklin Carrero-Martinez, Jennifer Saunders, and Emi Kameyama as a factual summary of what occurred at the workshop. The planning committee’s role was limited to planning the workshop. The statements made are those of the rapporteurs or individual workshop participants and do not necessarily represent the views of all workshop participants; the planning committee; or the National Academies of Sciences, Engineering, and Medicine.

REVIEWERS: To ensure that it meets institutional standards for quality and objectivity, this Proceedings of a Workshop—in Brief was reviewed in draft form by Aleksandra Berditchevskaia, Nesta; Kristi Ebi, University of Washington; and Sarah West, Stockholm Environment Institute. The review comments and draft manuscript remain confidential to protect the integrity of the process.

PLANNING COMMITTEE: Nebojsa Nakicenovic (Chair), The World In 2050; Aleksandra Berditchevskaia, Nesta; Luis Bettencourt, University of Chicago; Jeanne Holm, City of Los Angeles; Solomon Hsiang, University of California, Berkeley; and Barend Mons, Leiden University Medical Center.

Staff: Thomas Arrison, Director, Board on Research Data and Information (BRDI); Franklin Carrero-Martinez, Senior Director, Science and Technology for Sustainability Program (STS); Emi Kameyama, Program Officer, BRDI and STS; and Olivia Torbert, Senior Program Assistant, BRDI and STS.

SPONSORS: This workshop was supported by the National Academies George and Cynthia Mitchell Endowment for Sustainability Science and Schmidt Futures as part of BRDI core activities.

For additional information regarding the workshop, visit: www.nas.edu/sustainability and www.nas.edu/brdi.

Suggested citation: National Academies of Sciences, Engineering, and Medicine. 2022. Data-Informed Societies Achieving Sustainability: Tasks for the Global Scientific, Engineering, and Medical Communities: Proceedings of a Workshop—in Brief. Washington, DC: The National Academies Press. https://doi.org/10.17226/26513.

Policy and Global Affairs

Copyright 2022 by the National Academy of Sciences. All rights reserved.