United Nations Department of Economic and Social Affairs (UNDESA)
United Nations Department of Economic and Social Affairs (UNDESA)
Description of Activities on AI
DESA engaged with AI on a number of different fronts, from examining the use of AI by member states, \ utilizing AI in big data applications, to examining the impacts of AI on policy and the 2030 Agenda for Sustainable Development. The Department continues to support member states in providing policy analysis and research, sharing information, and capacity development assistance. The following are a selection of examples.
Project 1: United Nations E-Government Survey 2020
The DPIDG, in Chapter 6 “Towards Data-Centric E-Government” of the UN E-Government Survey 2020, has touched upon data and its use in AI. The need for data is nothing new but the ways in which data are created and used have changed dramatically in recent years, bolstered by the revolution in data technologies and the proliferation of applications of different types and forms of data, including small and big data, real-time data and geospatial data. The current COVID-19 pandemic also reinforces the centrality of data — how governments and businesses handle data, as it turns out, is a crucial part of their pandemic response. Learn more about open government data development, policy and institutional trends on government data sharing, exchange and interoperability, as well as data security, privacy and ethics; and recommendations on national data leadership and data governance framework.
Project 2: Big Data and Statistics using AI
The Statistics Division of UNDESA has been working on various initiatives related to artificial intelligence. This work is part of collaboration with national statistical institutes and other stakeholder groups under the umbrella of the UN Global Working Group (GWG) on Big Data for Official Statistics
Here are 3 concrete examples:
- FAO-UNSD project using satellite data and farm surveys to estimate crop statistics. The project aims to identify crops, map crop areas and estimate crop yield using satellite data and farm surveys. As AI approach it uses Supervised Machine Learning (random forests and support vector machines)
- Estimating Port Calls using AIS vessel tracking data. The project aims to identify ships which are entering and leaving a port (by vessel type) using AIS vessel tracking data. AIS data are real-time data of ship positioning. This is obtained as a global feed. As AI approach it uses Supervised Machine Learning (random forests).
- LinkedSDG, A demo app that automatically extracts key concepts related to sustainable development from text documents and links them to the most relevant sustainable development goals, targets, indicators and series. This uses Semantic Web technologies and ontologies, which is a subfield of AI and Computer Science research
Project 3: Research and engagement of scientists and engineers
TFM findings on the impacts of rapid technological change on the SDGs:
New and rapidly changing technologies, such as artificial intelligence, robotics and other automation technologies, biotechnologies, and nanotechnologies hold great promise for making accelerated progress towards the Sustainable Development Goals, but also pose formidable challenges in all of the SDG dimensions. Against this background, the UN General Assembly has called upon the UN Technology Facilitation Mechanism (TFM) in repeated resolutions to present their updated findings to the Annual Multi-stakeholder Forum on Science, Technology and Innovation for the SDGs (STI Forum). In response, “TFM findings on the impacts of rapid technological change on the SDGs” have been presented at every STI Forum since. The findings are crowdsourced from TFM partners and scientific and technological communities, through calls for inputs (including policy briefs and research papers), leveraging institutional networks, university partnerships and meetings. It is based mainly on volunteer work, and knowledge of technologies, sustainable development models and pathways has been built. In particular, since 2016, a series of UN expert group meetings on AI has provided a convergent series of general policy recommendations, upon which recommendations for specific issues elaborate. These efforts are ongoing. A series of lessons-learnt have been identified and important support provided to various reports. Present work in 2020 focuses on the environmental impacts of AI. A range of analytical and research material is being prepared and scientific data collected from contributors and volunteers will be publicly made available in the future. One of the key challenges has been the vast scope of the exercise, limited resources, and large expectations and the geopolitics of AI.
Exploring the impacts of new Internet applications and AI on the global energy system:
New Internet applications and especially AI technologies have become a rapidly increasing source of energy demand but have also greatly shaped the opportunities for smart and cleaner energy systems. This project reviews what is known and what might be potential policy responses to these trends in the future. The project addresses SDGs 7, 9 and 17. It draws on expert knowledge, volunteer work, and scientific networking. Key activities include desk study and expert surveys. The project is ongoing, and its expected output will be a research study and a database. A key challenge has been the identification of research work that exists in fragmented forms in various disciplines and both in academia and private sector. Hence, interdisciplinary expert surveys are key to their identification. Furthermore, a common technical terminology is needed.
IEEE/UN Event series
The UN and IEEE have recently partnered to organize a series of training and outreach events on technology, policy, ethics and engagement of AI and other new technologies. The Webinar series aims to engage engineers in the TFM process – to match the current level of engagement of scientists – and provide inputs on AI and related technologies for the TFM’s STI Forum. The initiative is ongoing and is expected to increase AI-related human capacity. Main partners are DESA, IEEE, the UNMGCY and TFM partners.
Long-term AI and technology scenarios for the SDGs
Long-term technology scenarios are routinely used to explore feasible technology pathways to tackle big global challenges, such as climate change and biodiversity. While an increasing number of them assume significant new opportunities due to AI, most of them do not make any effort to quantify these effects in both positive and negative terms. This new initiative aims to explicitly account for AI and potential future AI technology developments based on existing technology development data. It also provides inputs for the mandated discussions of long-term future scenarios and the impact of current trends in the high-level segment of ECOSOC each year (see, e.g., https://undocs.org/e/2020/60). The output will be a recurring research/study report and related AI scenario events with leading technology scenario analysis groups. The work draws on a variety of scenario models and aims to quantitatively account for technology change in general and AI deployment in particular. The initiative indirectly supports the achievement of all SDGs, but especially SDGs 7, 13, and 17. Ultimately, the initiative is expected to contribute to improving countries’ capacity through foresight analysis exercises and technology impact assessments and put them into a global context. This link between the global analysis and national-level level AI scenarios and roadmaps (where they exist) has been the primary challenge and opportunity at the same time.
Project 4: Capacity building tools
Guidebook on AI ethics for government and development practitioners:
While there are hundreds of publications and proposed AI ethics frameworks and codes of conduct by scientific and engineering communities, as well as an UNESCO initiative on AI ethics, little practical guidance exists for governments and development practitioners, especially guidance that is fully based on a balanced scientific and technological understanding. The guidebook aims to fill this gap. The initiative is ongoing and based on volunteer work by academics, practitioners and UN experts. The expected output is a collaborative guidebook developed by academics working on AI ethics with practical experience. Key SDGs addressed include SDGs 1, 2, 9, 16, and 17. Ultimately, it is expected to become a tool in the AI-related human capacity building. A key challenge is the translation of technical specificities into practical, easy understandable guidance for practitioners.
Guidebook to resources on AI strategies (supplement to the IATT Guidebook on STI roadmaps for the SDGs):
While there is an increasing number of AI strategies and an exponentially increasing number of publications on AI, government officials and development practitioners alike could benefit from a trusted, curated and annotated list of written resources on the various aspects of AI. The guidebook to resources on AI strategies is designed to do just that. It complements a more general IATT guidebook to STI roadmaps for the SDGs. It is currently under development by IATT and its partners. It focuses on SDG17 and aims to enhance technical infrastructure/support development of national and regional AI and digital strategies. The main challenge has been the informed selection of what constitute the most important yet accessible publications and resources on AI strategies.
Project 5: TFM online platform for sharing technology information
The TFM online platform (“2030 Connect”) was mandated to provide a single entry-point for technology information. It is an online platform for information on technologies and SDG knowledge. Its primary idea is to serve as a gateway to networks of curated SDG-related technologies and knowledge from UN and non-UN resources. DESA, OICT, UNCTAD, IATT and the TFM 10-Member Group have been main partners in its development (see https://tfm2030connect.un.org/). It has just recently been deployed and many partners and database providers have decided to partner and link their products to TFM connect. In this initial stage, the platform is looking for any interested partners to help growing the platform. Ultimately, it is expected to help strengthening technical infrastructure/support development of national and regional STI capacities for the SDGs in general and AI and digital strategies in particular. A key challenge continues to the extremely limited resources that constrain and prevent a more substantial operational component.
Project 6: UN Economist Network Report for the UN 75th anniversary
The UN Economist Network’s 2020 Report, Chapter 5: Emerging and frontier technologies (led by DESA and written in collaboration with other UN entities) discusses how near-universal digitalization, transformative technological breakthroughs and rapid diffusion of technologies are unleashing structural shifts that are long term and irreversible, with far-reaching consequences. The chapter has comprehensive discussions on how advances in AI could affect labour market outcomes, help policymakers to better understand urbanization, and improve provision of public services. It notes the potential of AI in improving the efficiency of energy systems, but also highlights that AI technologies are significant energy consumers and further application of such technologies could significantly ramp up energy demand and emissions. The chapter calls for data policies that protect individuals’ rights, foster open-data policies, create standard for the interoperability of data functions, and advance skills relevant for the data economy.
Project 7: Frontier Technology Issues
The Economic Analysis and Policy Division produces the Frontier Technology Issues, which are a series of regular reports that delve into specific aspects of a new technology, including its associated development challenges and policy implications. In 2020, two reports were produced. The first one, titled “Does the sharing economy share or concentrate”, discusses the effects of the expansion of the sharing economy – underpinned by AI technologies – on income inequality, which vary depending on each country’s development conditions and policies. The paper discusses policy areas where concerted actions have to be taken to ensure that the sharing economy would not worsen inequality in an already highly unequal world. The second report, titled “Can digital technologies put us back on the path to achieve the SDGs?”, shows how countries are using technologies, including AI, in response to the pandemic, to create decent work (SDG 8), improve health services and outcomes (SDG 3), and promote education and learning (SDG 4). It also warns that technology is widening disparities if unchecked, and those on the wrong side of the digital divide risk being left further behind.
Challenges and Opportunities
With the vast scope of AI and large expectations of its potential, resource constraints and limitations in hiring or developing human resources with the right skills mix, there can be significant challenges in AI projects. Measuring progress and impact is also a key challenge.