Policy Brief No. 225 — February 2026
AI’s Global South Pivot: Equity, Ethics and Ecology
Deepak Maheshwari
Key Points
→ Investment in core technologies and infrastructure for artificial intelligence (AI) continue to be concentrated within the Global North, even as the Global South — home to 88 percent of humanity — remains more vulnerable to risks.
→ Structural exclusion perpetuates the Global South’s post-colonial dependency as “rule-takers” who provide training data, deployment sites, natural minerals and other critical elements for this transformative technology, albeit without adequate safeguards or commensurate social and economic gains. To remedy this “wicked problem,” AI governance frameworks must be predicated on the trinity of equity, ethics and ecological sustainability.
→ This policy brief identifies opportunities for inclusive, effective and proportionate participation by the Global South in AI governance, outlining the systemic and power dynamics at play, and offering actionable recommendations for the short, medium and long terms.
Introduction
The impact of the 2022 launch of ChatGPT (OpenAI 2022), a generative AI tool,1 calls to mind that of Netscape Navigator’s debut in 1994. Both products democratized access to transformative technologies with seemingly magical capabilities. ChatGPT enabled natural language interactions, making it user-friendly for the average user, just as the Netscape browser had made the internet broadly accessible, and both were offered free of charge, at least initially.
However, internet access had increased significantly across the Global South by the time ChatGPT was launched, whereas three decades earlier, when Netscape Navigator was released, internet services were mostly concentrated within the Global North.2 All the same, even in 2026, AI infrastructure, technologies and investment continue to be concentrated within the Global
1 IBM’s Deep Blue and Watson, followed by Google’s AlphaGo debuting in 1995, 2004 and 2016, respectively did generate excitement, but they did not lead to mass adoption.
2 Neither the Global South nor the Global North is a monolithic or homogeneous region. The constituents of the Global South are also sometimes referred to as low- and middle-income countries.
About the Author
With more than three decades of experience, Deepak Maheshwari is a thought leader with sharp insights. He has a keen interest in the interplay of public policy with technological innovation and socio-economic development through an interdisciplinary lens.
Currently, he is a senior policy advisor at the Centre for Social and Economic Progress, a consultant with the Indian Council for Research on International Economic Relations and an advisor at the Indicus Centre for Financial Inclusion. He has also been affiliated with the Consumer Unity & Trust Society, the Center for the Digital Future, the Public Affairs Forum of India and Palo Alto Networks. An oft-invited speaker and columnist, his views are widely sought, published and cited for their clarity and holistic perspective.
In addition to serving on the government committees on artificial intelligence and accessibility, he has volunteered as global chair of the IEEE Internet Initiative, secretary of the Internet Service Providers Association of India and advisory board member of the IDEA Telecom Centre of Excellence at the Indian Institute of Management, Ahmedabad. Earlier, he led the public policy function at Microsoft, Mastercard, Symantec and Sify, covering India, South Asia, and the Association of Southeast Asian Nations and China regions.
A strong believer in public-private partnerships, Deepak co-founded the National Internet Exchange of India and the ITU-APT Foundation of India. He holds an engineering degree from the Indian Institutes of Technology as well as a law degree.
North barring a few Global South countries such as China, India and Brazil.
Meanwhile, use cases of AI have been multiplying — with applications from improving health-care outcomes and educational simulations, to weather forecasting and developing climate solutions. It can also help in improving delivery of public services, especially relevant in the Global South (Zugravu et al. 2024).
However, there are numerous concerns as well, many already evident. These range from AI technologies’ use to power misinformation campaigns and cyberattacks targeting critical infrastructures, the viral creation and distribution of child sexual abuse material, Orwellian surveillance, systemic exclusion and isolation of vulnerable communities and the spectre of job losses en masse. Other concerns include dominant technology companies’ alleged antitrust behaviour, invasion of individuals’ privacy, inherent or induced bias in product design, and adverse ecological impact.3
While the Global North continues to accrue most of AI’s benefits, the Global South remains disproportionately exposed to its risks. The reason is that the AI systems have often been designed and deployed without factoring in the local contexts and realities of the Global South, and these risks are further compounded due to its larger population base. This imbalance limits the dispersal of AI’s benefits within the Global South but also leads to the suboptimal evolution and use of AI at large — adversely impacting even the Global North in the long term.
Moreover, despite holding large deposits of natural resources needed for extraction of critical minerals, the Global South gains little economic value from AI but bears high ecological costs (International Renewable Energy Agency 2024).
Meanwhile, the contours of AI governance continue to be shaped
3 See www.iea.org/data-and-statistics/data-tools/energy-and-aiobservatory?tab=Energy+for+AI.
by the Global North, with the Global South playing a marginal or symbolic role at best.
By identifying systemic and power dynamics at play with respect to AI governance, this policy brief is an endeavour to recommend actionable pathways to ensure that the Global South emerges as an equitable contributor, shaper and leader in the realm of AI governance rather than remaining a passive follower. It begins with an analysis of AI’s constituents and contexts as well as its promises and perils.
Contextualizing Constituents of AI
Since the term “artificial intelligence” was coined in 1956,4 its scope has seen multiple evolutions from machine learning to generative and agentic AI (Karjian 2024). The output of an AI system — whether deterministic, probabilistic or even a hybrid thereof — entails transformation and processing of inputs by orchestrating specialized constituents: data; algorithms; model architectures; training processes; inference engines; computing infrastructures; evaluation metrics; deployment and integration; and security, law and ethics. Despite being unique, each of these constituents is often interdependent on others, even if spatially dispersed around the globe.
As data is both the key input and the key output of AI, significant investments are under way toward its generation, collection, curation and correlation. Obviously, within the policy discourse, there is a lot of focus on data — whether personal or non-personal, natural or synthetic. In addition, there is growing demand for seeking and ensuring fairness, accountability and transparency in algorithms, training processes and inference engines. These trends warrant a closer look at overall AI governance.
AI Governance: A Wicked Problem
As AI depends on multiple factors and impacts multiple interdependent constituents and domains, its governance is a “wicked problem” (Rittel and Webber 1973), because the challenges are deeply entangled, dynamically evolving and value-laden. Systems are often opaque and non-stationary even as every solution changes incentives and creates unforeseeable and unintended risks, calling for trade-offs (Gurumurthy and Chami 2019). For example, one must simultaneously navigate multiple potentially conflicting objectives: fairness versus speed, safety versus innovation, transnational cooperation versus national sovereignty, and ecological sustainability versus developmental demands. Boundaries blur across ethics, security, labour, environment and geopolitics, while measurement mechanisms remain contested. Governance, therefore, must be dynamic, multi-dimensional and inclusive — able to adapt in real time and reconcile such challenges without any pretence of permanent resolution.
A plethora of frameworks for AI governance have been proposed or codified5 at an accelerated pace and with increasing levels of complexity and details, with some of the most prominent ones curated by the Global AI Ethics and Governance Observatory of the United Nations Educational, Scientific and Cultural Organization (UNESCO).6
The proponents include multilateral organizations, for example, the United Nations and its constituent entities such as the International Telecommunication Union (ITU) and UNESCO (UNESCO 2022); the Organisation for Economic Co-operation and Development (OECD); the World Bank; the European Union, the Council of Europe and the Group of Twenty (G20); national governments such as those of the United States (White House 2022), the United Kingdom (Department for Science, Innovation and Technology 2023), China (Zhang 2023) and India (Ministry of Electronics and Information Technology [MeitY] 2025; National Institute for Transforming India [NITI] Aayog 2021); and
5 See Appendix A for key AI governance frameworks.
4 See https://home.dartmouth.edu/about/artificial-intelligence-ai-coineddartmouth.
6 See www.unesco.org/ethics-ai/en; see Appendix B for other repositories.
standards bodies such as the International Organization for Standardization, the International Electrotechnical Commission and the Institute of Electrical and Electronics Engineers (IEEE);7 civil society organizations such as The Future Society (Jeanmaire et al. 2025); and even companies and coalitions.
However, most of the frameworks are voluntary, often forsaking enforceability for agility, as models, technologies and markets are in a constant flux of evolution. Even where hard law exists, such as the EU AI Act, cross-border coordination lags, and the absence of shared definitions — of “high-risk,” “foundation models” or “responsible use” — keeps consensus out of reach, with limited attention to ecological impact.
Such divergence is structural and not accidental. Some jurisdictions anchor governance in human rights and individual liberties; others focus on strategic competitiveness and geopolitical context. The European Union and Brazil lean toward broad, uniform rules, while the United States and the United Kingdom favour sectoral specificity and experimental sandboxes. Recently, India has adopted an approach that “balances AI innovation with accountability, and progress with safety” (MeitY 2025, 1).
These choices reflect distinct legal systems and business interests. They set path-dependent trajectories, making subsequent convergence and alignment even harder, particularly for countries in the Global South, with rising costs for compliance and switching across frameworks.
Power concentration further compounds the wickedness of the problem. Frontier capabilities, computing infrastructure, data and talent are clustered within a handful of firms and research hubs within the Global North. Attempts to stall the race, such as the 2023 “pause letter” (Future of Life Institute 2023), failed spectacularly, exposing collective action dilemmas, such as low enforceability and strong first-mover advantage; however, the letter did succeed in mainstreaming the conversation around AI governance (Heikkilä 2023). Without credible mechanisms for oversight, liability and access,
such actions risk legitimizing and perpetuating existing asymmetries rather than resolving them.
With limited resources and institutional capacities, the Global South countries face outsized impacts such as labour displacement due to automation, surveillance risks due to black-box technologies, and ecological imbalance due to extensive mining and high demand of energy for the AI ecosystem. Moreover, users in the Global South often end up using complex models and technologies that were developed without considering their local contexts or concerns, such as low-bandwidth networks, predominance of non-English languages and energy deficits even for household supplies. Hence, it is useful to holistically examine the structural and historical exclusion of the Global South in global governance in general and AI governance in particular.
Systemic Exclusion of the Global South
Exclusion of the Global South in shaping global norms is not a new phenomenon, however. For example, to mitigate adverse impacts of climate change, the Global South must bear a disproportionate burden of commitments even as the Global North has all but retracted from its legally binding commitments to provide finance and technology under the Paris Agreement.8 Likewise, the models for financial services, such as for payments from the Global North, are misfits for the majority of the population, and the adoption of certain technologies may lead to vendor lock-in.
In a perpetuation of the post-colonial dependency, countries in the Global South are often considered and even made to believe and behave as “rule-takers,” a pattern visible even in the case of AI governance. Countries within the Global South are essentially seen as providers of data, both personal and non-personal, as well as critical sites for experimentation and deployment of AI with little accountability.
7 See https://standards.ieee.org/initiatives/autonomous-intelligencesystems/.
8 United Nations Framework Convention on Climate Change, Adoption of the Paris Agreement, 12 December 2015, Dec CP.21, 21st Sess, UN Doc FCCC/CP/2015/L.9. See the status of the treaty at https://treaties.un.org/pages/ViewDetails.aspx?src=TREATY&mtdsg_ no=XXVII-7-d&chapter=27&clang=_en.
However, they continue to face the dual challenges of, firstly, lacking financial and technical resources and, secondly, being deprived of an opportunity to shape the global governance norms. This is even more concerning since the 134 countries9 in the Global South account for 88 percent of the global population (Mahbubani 2024) and 90 percent of the population under 25 years of age,10 and the emerging markets are expected to contribute 65 percent of gobal economic growth by 2035 (Perez-Goropze, Cardenas and Tesfay 2024). Consider that less than two percent of global AI investment is in Africa — even as it is home to 18 percent of humanity — and this inequity becomes even more stark, compounded by electricity access gaps affecting 750 million people globally, with 80 percent in Sub-Saharan Africa alone (Cozzi et al. 2024).
Moreover, the Global South countries are not only nudged to adopt regulations that are unsuitable to their local contexts, but they are also without the requisite state capacity for enforcement. For example, while the European Union’s General Data Protection Regulation spurred many countries in the Global South to enact privacy laws, it strained limited state capacity, with fines and compliance costs hitting domestic small and medium enterprises. The EU AI Act’s extraterritorial effects similarly pushes risk-based approaches within the Global South but risks overregulation.
Such “imported” frameworks, technologies or models often do not factor in the enormous diversity within the Global South — and its equally diverse range of challenges, from those of the small island nation-states such as Fiji and the Maldives to the under-representation of Indian and African languages and scripts, dialects and accents due to the predominant bias of AI models toward the English language and the Roman script. The ecological dilemma is also acute. Despite having rich deposits of critical minerals such as lithium, cobalt and copper — essential for AI hardware — countries in the Global South, particularly in Africa, export these minerals cheaply but then bear high prices for chips embedded with the same critical minerals, imported from the Global North, China, Japan and Korea.
Several efforts, endogenous as well as exogenous, are under way to enable and enhance effective participation of the Global South in AI governance in response to such systemic inequities.
Some Silver Linings
There is growing momentum for including the Global South ab initio in discussions and platforms, advisory bodies and standard-setting bodies, while also providing targeted support for investment in policy development and digital infrastructure rollout, and enhancing digital literacy, innovative use cases and a focus on ecological sustainability, especially in Africa and South Asia.
Enshrining inclusive, multilateral AI governance, the UN Global Digital Compact urged for broad-based consultations, elevating Global South priorities in policy dialogues (United Nations 2024). Similarly, the Global Dialogue on Artificial Intelligence Governance during the eightieth session of the UN General Assembly underscored three core themes: AI governance must be inclusive, science-based11 and rights-respecting; it must promote interoperability, open innovation and shared capacity building; and it must remain human-centric (United Nations 2025).
The World Bank has also urged the Global South to adopt robust AI governance (World Bank Group 2024) and it supports flexible and adaptable national AI strategies focused on AI for public good (World Bank Group 2025).
The G20 (G20 Leaders 2023) and the Global Partnership on Artificial Intelligence (GPAI) (GPAI Ministers 2024) championed the need for equitable standards and non-weaponization of AI, with India playing a leading role, in line with its strategy focused on “AI for All” (NITI Aayog 2018), and leveraging the complimentary relationship between digital public infrastructure (DPI) and AI (Arakali 2025), that can be succinctly articulated as “DPI democratizes AI and AI supercharges DPIs” — an idea also endorsed by the World Economic Forum (2025), with calls for eco-friendly AI designs.
9 See https://worldpopulationreview.com/country-rankings/global-southcountries.
10 See https://lausanne.org/report/demographics/youth.
11 The 40-member Independent International Scientific Panel on AI being set up by the United Nations for “advancing global, evidencebased understanding of artificial intelligence (AI).” See www.un.org/ independent-international-scientific-panel-ai/en/open-call.
The focuses of the annual AI summits indicate similar shifts. The summit launched in 2023 as the AI Safety Summit in the United Kingdom with discussion surrounding particular safety risks of frontier AI technologies and how addressing them would require global cooperation. Co-hosted by the United Kingdom and South Korea in 2024, the AI Seoul Summit considered the interrelated goals of Al safety, innovation and inclusivity. In 2025, France hosted the AI Action Summit in Paris to discuss the global governance and development of sustainable AI that served the public interest and allowed shared progress.
In 2026, the AI Impact Summit will be held in Delhi — India will be the first Global South country to host. Focused on the principles of people, planet and progress, the summit’s working groups are around inclusion, economic growth and social good, as well as on democratizing AI resources. Considering the growing energy demand for the data centres12 and the challenges of physical resources, including critical minerals, the 2026 summit will focus on “advancing resourceconscious and adaptable AI systems that operate effectively in low-resource environments without compromising accessibility or impact.”13
Notably, “Building Governance Together,” the theme of the Internet Governance Forum (IGF) 2025 in Norway, was also reflective of the move toward a more inclusive and equitable global digital governance framework.14 The Inter-Parliamentary Union (IPU) insists on inclusion, particularly “the need for the active presence of the Global South” (IPU 2024, paragraph 3). Likewise, the Global AI Summit on Africa underscored the guiding principles of inclusivity and diversity while also calling for the optimization of compute resources with minimal carbon footprint.15
Another significant development is that the policy makers are increasingly adopting and promoting open technologies that are freely available to use, study, modify and share. These include opensource software, DPI (Maheshwari 2025a), O-RAN (open radio access networks) for 5G and 6G
12 See www.iea.org/data-and-statistics/data-tools/energy-and-aiobservatory?tab=Energy+for+AI.
13 See https://impact.indiaai.gov.in/working-groups/resilience-innovationefficiency.
14 See www.igf2025.no/.
15 See https://c4ir.rw/docs/Africa-Declaration-on-Artificial-Intelligence.pdf.
mobile communications, and, in the context of AI, architecture, weights and even training codes. This augurs well for the Global South in mitigating the risk of vendor lock-in and keeping costs under check, whether for AI integration in 5G and 6G networks or for the use of AI for public services such as health care and education.
Other efforts are focused on improving the efficacy of regional coordination or maximizing the impact through horizontal transfers within the Global South, at times but not always aided by multilateral organizations or countries in the Global North. As a result, there are regional AI observatories such as the MENA [Middle East and North Africa] Observatory on Responsible AI,16 South-South cooperation and even trilateral cooperations (United Nations Industrial Development Organization 2025) aiming, for example, to build local processing for critical minerals.17
China — the second-largest economy in the world and a global leader in AI research and deployment — has also emerged as the leading manufacturing hub for hardware within the global supply chains. In that sense, it could be considered as an outlier among the Global South countries.
Since most AI systems are modelled around data, concepts and context prevailing in the Global North, at times these may be misfits or misaligned in the Global South, or simply unaffordable, unless they are duly tweaked or developed from the ground up and factor in the local context, culture, computing infrastructure and capabilities, including electricity constraints.
The following examples demonstrate the challenges of retrofitting AI solutions without factoring in the host environment:
→ In 2018, a study of three leading AI facial-analysis applications showed an error rate of 34.7 percent when analyzing dark-skinned females’ faces compared to 0.8 percent for light-skinned males’ faces (Hardesty 2018).18
16 See https://menaobservatory.ai/en/home.
17 See www.africa-press.net/rwanda/policy/critical-minerals-highlighted-atrwanda-mining-week-2025.
18 Though not directly about AI governance, this case highlights the need for inclusive design and deployment to mitigate bias, even if unintentionally introduced.
→ Challenges with under-represented languages, phonetic scripts and semantic differences from Anglo-Saxon languages necessitate further advances in sentiment analysis and text normalization (Syauqi and Wibawa 2025). For example, English-language dominance within machine translation often loses the contextual meaning within Nigerian languages (Olojo et al. 2025).
→ Introduction of autonomous vehicles designed for advanced countries may lead to unmitigated disasters in the Global South where roadways have significant differences: numerous potholes; lane driving as an exception; shared use by pedestrians, vendors and cattle; and many different types of vehicles, including bicycles, motorcycles, tuk-tuks, cars, trucks and tractors (Sirikhan 2022). Add a rattling bullock cart into the mix and the algorithm would be further flummoxed.
Deploying more computing power and making data networks faster are not pragmatic approaches. Instead, innovative approaches are needed to solve such challenges with local insights leveraging these variations as features, not bugs. Here are a couple of examples:
→ Enabling real-time text, speech and voice translation, Bhashini19 uses English as the pivot for translation with just 22 language-pair models instead of 231 pairs of one-on-one translation corpora across the 22 official languages in India.
→ Omni-directional microphones at intersections harness honking tendency to detect traffic slowdowns in India with more than 85 percent accuracy, offering a low-cost alternative to expensive and complex cameras and sensors that fail in unmarked, mixed-traffic lanes.
Clearly, solutions need not be complex, costly or demanding extreme computing power or bandwidth. However, they must be contextually relevant and frugal yet smart, prioritizing ecological sustainability.
At the same time, open-source software and DPI democratize technology adoption with potential cost savings and the ability to foster innovation. While international partnerships are
helping to fund AI innovation hubs and startup ecosystems within the Global South, civil society can help traverse the path from grassroots campaigns to policy advocacy with local evidence, advocating for sustainable mineral economies.
Lessons and Recommendations
Constraints experienced by Global South countries include limited awareness of AI; inadequate digital inclusion, funding, technical infrastructure and capabilities; and a lack of opportunities to participate in appropriate fora and decisionmaking institutions. The implications of these constraints are grave, for both the Global South and the Global North. The repercussions may manifest in different ways — widening inequality, sudden and mass labour displacement and largescale migration, leading to social and political instability, economic deceleration and even largescale conflicts. Since addressing these imbalances is more about global resilience than mere fairness, exclusion of the Global South from AI governance risks undermining trust in the very systems AI governance frameworks seek to regulate, especially amid ecological strains such as critical mineral exploitation and energy shortages.
History shows that the exclusion mainly emanates from asymmetric power structures even as unintentional exclusion may arise from language barriers, lack of opportunity or less diverse training data for models being developed in the Global North. For example, while the World Summit on the Information Society 2003 established multi-stakeholder principles, power imbalances persist. The IGF’s two decades of experience demonstrates success in convening and facilitating dialogues but also failure in securing binding outcomes or commitments, lessons relevant for sustainable AI governance.
Useful for every country, but especially so for those in the Global South, here are specific actionable recommendations for AI governance — aligned with the United Nations’ Sustainable Development Goals (SDGs) and grounded on the three foundational principles of equity, ethics and ecological sustainability.
19 See https://bhashini.gov.in.
Short Term (Up to 2027)
Global South countries should create national strategies for digital inclusion and AI adoption by:
→ aligning with the SDGs;
→ using multi-dimensional metrics such as the Government AI Readiness Index (Nettel et al. 2024) and the Global Index on Responsible AI;20
→ redesigning institutional framework and policy architecture to leverage increasing convergence across telecommunications, information technology and broadcasting (Maheshwari and Sharma 2025);
→ prioritizing resource efficiency as well as open technologies, adoption of DPI, state capacity, and leveraging of public-private partnerships (Maheshwari 2025b);
→ championing the pivotal role of the Global South in AI governance during the India AI Impact Summit 2026; and
→ highlighting the need for building state capacity for AI governance in the World Bank’s World Development Report 2026.
Medium Term (Up to 2030)
Global South countries should ensure that:
→ AI infrastructure rollout is accompanied by initiatives to empower communities for innovative AI usage such as preservation of traditional knowledge and support for local languages;
→ adoption of interoperable trust architectures (such as digital identity, payment and consent protocols) is deployed with external audit and oversight mechanisms;
→ South-South and triangular cooperation frameworks are used for speed, scale and sustainability; and
→ cybersecurity strategy and data governance frameworks are updated and adapted for the AI era, while also considering their specific needs, resources and capabilities.
Long Term (Up to 2035)
Global South countries should:
→ seek, secure and sustain proportionate representation in transnational and global fora;
→ strategically partner with willing countries and regions from both the Global South and the Global North for AI deployment as well as for governance, albeit without compromising their national sovereignty;
→ inform, influence and impact the emerging frameworks for AI governance at national, regional and international fora; and
→ inform, influence and internalize flexible and ethical AI frameworks.
20 See Appendix C for key AI readiness and governance indices that can be tools for global benchmarking.
Appendix A: Key Frameworks for AI Governance
EU AI Act
European Commission
Framework Convention on AI Council of Europe
OECD AI Principles
UNESCO’s Recommendation on the Ethics of Artificial Intelligence
United States, “America’s AI Action Plan”
of Science and Technology Policy
United Kingdom’s policy paper “A proinnovation approach to AI regulation” Department for Science, Innovation and Technology
Singapore’s Model AI Governance Framework
India AI Governance Guidelines
China’s Interim Measures for the Administration of Humanized Interactive Services
Infocomm Media Development Authority
Cyberspace Administration of China; Ministry of Industry and Information Technology
since 2024
treaty
protect fundamental rights through riskbased regulation of AI
ensure AI life cycle is consistent with human rights, democracy and the rule of law
promote innovation and human-centric AI
safeguard human rights and cultural diversity
achieve and maintain global leadership in AI security AI
development To promote innovation with proportionate regulatory measures for risk management
provide practical ethical guidelines for industry
prevent harms from AI-related human vulnerabilities through contextual technical measures
Canada’s Directive on Automated Decision-Making
Treasury Board of Canada Secretariat 2019 Mandatory for federal agencies using AI for administrative decison making
Japan’s AI Governance Guidelines Cabinet Office of Japan
“Ten AI Governance Priorities”
IEEE’s Ethically Aligned Design: A Vision for Prioritizing Human Well-being with Autonomous and Intelligent Systems (1st ed.)
The Future Society 2025 Civil society–driven road map
framework
To ensure that AI use for automated decision making by government agencies is transparent and accountable
To foster trustworthy AI and societal acceptance
To urge safeguarding of humanity through enforceable rules, audits, crisis response and inclusive oversight
To embed ethical principles into design, development and deployment of technology
Appendix B: Regularly Updated Repositories Tracking AI Governance Frameworks
Tracking and comprehending such diverse frameworks is nearly impossible, let alone complying with them. While more fragmentation in the near term is almost certain, some sort of convergence or consolidation is quite likely in the medium term, provided all stakeholders join hands in earnest — including those across the Global South. All the same, some of the important repositories are listed here.
Resource Description Link
OECD AI Policy Observatory
International Association of Privacy Professionals’ Global AI Law and Policy Tracker
UNESCO Global AI Ethics and Governance Observatory
Interactive dashboard of national AI policies, frameworks and indicators
Country-by-country breakdown of AI-related laws and proposals
Map of global AI policies, ethics guidelines and regulatory instruments
Awesome AI Ethics (GitHub)
AI Governance Library
AI Governance Observatory
Community-curated list of ethics guidelines, frameworks and tools in AI
Collection of white papers, templates and checklists for designing and implementing AI governance
Collaborative hub for AI governance professionals, researchers and civil society actors
https://oecd.ai
https://iapp.org/resources/article/ global-ai-legislation-tracker
www.unesco.org/ethics-ai/en
https://github.com/ awesomelistsio/ awesome-ai-ethics
www.aigl.blog/
https://ai-governanceobservatory.org
Appendix C: Global Benchmarking and Indices
Global benchmarking using AI readiness and governance indices can be a powerful tool for both stakeholders in the Global South and international partners. It helps identify gaps, informs appropriate policy and directs investment, supports regional cooperation and strengthens global cohesion.
Some key indices and frameworks are listed in the table.
Index or Framework
Government AI
Readiness Index
Agency or Initiative
Oxford Insights and the International Development Research Centre (IDRC)
Focus Areas
Public sector AI readiness across 188 countries
Global Index on Responsible AI
Global Center for AI Governance
Human rights–based AI governance in 138 countries
AI Readiness Framework ITU and AI for Good Standardized readiness across sectors (health, agriculture, disaster resilience)
AI Index Report
Stanford University Human-Centered AI
Global South Relevance
Widely used by the United Nations, the G20 and UNESCO; good coverage of low- and middleincome countries
Designed for inclusive benchmarking; supported by Global Affairs Canada and IDRC
Tailored case studies from Asia, Africa and Latin America
Technical progress, policy and societal impact
Tracks regional divides and Global South optimism
Acronyms and Abbreviations
AI artificial intelligence
DPI digital public infrastructure
G20 Group of Twenty
GPAI Global Partnership on Artificial Intelligence
IDRC International Development Research Centre
IEEE Institute of Electrical and Electronics Engineers
IGF Internet Governance Forum
IPU Inter-Parliamentary Union
ITU International Telecommunication Union
MeitY Ministry of Electronics and Information Technology
NITI National Institute for Transforming India
OECD Organisation for Economic Co-operation and Development
SDGs Sustainable Development Goals
UNESCO United Nations Educational, Scientific and Cultural Organization
Works Cited
Arakali, Harichandan. 2025. “AI can turbocharge Digital Public Infrastructure and unleash benefits at scale.” Forbes India, February 24. www.forbesindia.com/article/leadership/ ai-can-turbocharge-digital-public-infrastructure-andunleash-benefits-at-scale-nandan-nilekani/95387/1.
Cozzi, Laura, Katarina Malmgren, Nouhoun Diarra, Gianluca Tonolo and Daniel Wetzel. 2024. “Electricity access continues to improve in 2024 — after first global setback in decades.” International Energy Agency, November 15. www.iea.org/commentaries/electricity-access-continuesto-improve-in-2024-after-first-global-setback-in-decades.
Department for Science, Innovation and Technology. 2023. “AI regulation: a pro-innovation approach.” Policy paper, March 29. London, UK: Government of the United Kingdom. www.gov.uk/government/publications/ ai-regulation-a-pro-innovation-approach.
Future of Life Institute. 2023. “Pause Giant AI Experiments: An Open Letter.” March 22. https://futureoflife.org/ open-letter/pause-giant-ai-experiments/.
G20 Leaders. 2023. “G20 New Delhi Leaders’ Declaration.” September 9–10. www.mea.gov.in/Images/CPV/ G20-New-Delhi-Leaders-Declaration.pdf.
GPAI Ministers. 2024. “2024 GPAI New Delhi Declaration.” July. https://wp.oecd.ai/app/uploads/2025/01/ gpai-new-delhi-declaration-2024.pdf.
Gurumurthy, Anita and Nandini Chami. 2019. The Wicked Problem of AI Governance. New Delhi, India: Friedrich-Ebert-Stiftung. https://itforchange.net/sites/default/files/1625/The%20 Wicked%20Problem%20ofAI%20Governance.pdf.
Hardesty, Larry. 2018. “Study finds gender and skin-type bias in commercial artificial-intelligence systems.” MIT News, February 11. https://news.mit.edu/2018/study-findsgender-skin-type-bias-artificial-intelligence-systems-0212.
Heikkilä, Melissa. 2023 “What’s changed since the ‘pause AI’ letter six months ago?” MIT Technology Review, September 26. www.technologyreview.com/2023/09/26/1080299/ six-months-on-from-the-pause-letter/.
International Renewable Energy Agency. 2024. Geopolitics of the Energy Transition: Energy Security. Abu Dhabi, United Arab Emirates: International Renewable Energy Agency. www.irena.org/Publications/2024/Apr/ Geopolitics-of-the-energy-transition-Energy-security.
IPU. 2024. “Resolution on impact of artificial intelligence (AI) on democracy, human rights and the rule of law.” 149th IPU Assembly, Geneva, October 13–17. www.ipu.org/impact/ democracy-and-strong-parliaments/artificial-intelligence.
Jeanmaire, Caroline, Jessica Galissair, Tereza Zoumpalova and Niki Iliadis. 2025. “Ten AI Governance Priorities: Survey of 44 Civil Society Organizations.” The Future Society, April 2. https://thefuturesociety.org/cso-ai-governance-priorities/.
Karjian, Ron. 2024. “The history of artificial intelligence: Complete AI timeline.” TechTarget, September 24. www.techtarget.com/searchenterpriseai/tip/ The-history-of-artificial-intelligence-Complete-AI-timeline.
Mahbubani, Kishore. 2024. “Measuring the power of the Global South.” The World Today, Chatham House, February 2. www.chathamhouse.org/publications/ the-world-today/2024-02/measuring-power-global-south.
Maheshwari, Deepak. 2025a. Digital Inclusion: Policy Pathways for Public-Private Partnerships. Policy Brief No. 208. Waterloo, ON: CIGI. www.cigionline.org/publications/ digital-inclusion-policy-pathways-forpublic-private-partnerships/.
———. 2025b. “A Novel Institutional Architecture for Digital Public Infrastructure.” Opinion, Centre for International Governance Innovation, February 14. www.cigionline.org/articles/ a-novel-institutional-architecture-fordigital-public-infrastructure/.
Maheshwari, Deepak and Bhavna Sharma. 2025. Governing Digital India: A Report on Institutions and Instruments. New Delhi, India: Centre for Social and Economic Progress. https://csep.org/wp-content/ uploads/2025/10/GOVERNING-DIGITAL-INDIA.pdf.
MeitY. 2025. India AI Governance Guidelines: Enabling Safe and Trusted AI Innovation. Government of India, November 5. https://indiaai.s3.ap-south-1. amazonaws.com/docs/guidelines-governance.pdf.
Nettel, Pablo Fuentes, Emma Hankins, Richard Stirling, Giulia Cirri, Gonzalo Grau, Sulamaan Rahim and Eddie Crampton. 2024. Government AI Readiness Index 2024. Oxford Insights. https://oxfordinsights.com/ ai-readiness/ai-readiness-index/.
NITI Aayog. 2018. National Strategy for Artificial Intelligence June. www.niti.gov.in/sites/default/files/2023-03/ National-Strategy-for-Artificial-Intelligence.pdf.
———. 2021. Approach Document for India, Part 1 — Principles for Responsible AI. February. www.niti.gov.in/sites/ default/files/2021-02/Responsible-AI-22022021.pdf.
OpenAI. 2022. “Introducing ChatGPT.” OpenAI, November 30. https://openai.com/index/chatgpt/.
Olojo, Seyi, Janina Zakrzewski, Andrew Smart, Erin van Liemt, Milagros Miceli, Amber Ebinama and Lameck Mbangula Amugongo. 2025. “Lost in Machine Translation: The Sociocultural Implications of Language Technologies in Nigeria.” In FAccT ’25: Proceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency, 2249–59. New York, NY: Association for Computing Machinery. https://doi.org/10.1145/3715275.3732153.
Perez-Goropze, Jose, Carlos Cardenas and Natznet Tesfay. 2025. “Emerging Markets: A Decisive Decade.” S&P Global, October 16. www.spglobal.com/en/ research-insights/special-reports/look-forward/ emerging-markets-a-decisive-decade.
Rittel, Horst W. J. and Melvin M. Webber. 1973. “Dilemmas in a General Theory of Planning.” Policy Sciences 4 (2): 155–69. www.sympoetic.net/Managing_Complexity/complexity_ files/1973 Rittel and Webber Wicked Problems.pdf.
Sirikhan, Kulacha. 2022. “Potentials and Challenges of The Connected Autonomous Shared Electric Vehicle (CASE) from Urban Geography Perspective in Southeast Asia Mega-Urban Regions.” Preprint, HAL Science, March 14. March. https://hal.science/hal-03607999v1.
Syauqi, A’la and Aji Prasetya Wibawa. 2025. “Advances in machine transliteration methods, limitations, challenges, applications and future directions.” Natural Language Processing Journal 11 (June): 100158. https://doi.org/10.1016/j.nlp.2025.100158.
The White House. 2022. Blueprint for an AI Bill of Rights: Making Automated Systems Work for the American People Washington, DC: Office of Science and Technology Policy. https://bidenwhitehouse.archives.gov/wp-content/ uploads/2022/10/Blueprint-for-an-AI-Bill-of-Rights.pdf.
UNESCO. 2022. Recommendation on the Ethics of Artificial Intelligence. Paris, France: UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000381137.
United Nations. 2024. Pact for the Future, Global Digital Compact and Declaration on Future Generations. Summit of the Future Outcome Documents, September. www.un.org/sites/un2. un.org/files/sotf-pact_for_the_future_adopted.pdf.
———. 2025. “Global Dialogue on Artificial Intelligence Offers Platform to Build Safe Systems, Secretary-General Says at Launch.” Press release, September 25. https://press.un.org/en/2025/sgsm22839.doc.htm.
United Nations Industrial Development Organization. 2025. “Leveraging AI for Sustainable Development through South-South and Triangular Industrial Cooperation.” Concept Note, May 16. https://unsouthsouth.org/ wp-content/uploads/2025/05/UNIDO-SSTIC-Side-Eventat-the-22nd-HLC-on-SSC-Concept-Note-v20250516.pdf.
World Bank Group. 2024. Global Trends in AI Governance: Evolving Country Approaches. Washington, DC: International Bank for Reconstruction and Development. https://documents1.worldbank.org/ curated/en/099120224205026271/pdf/ P1786161ad76ca0ae1ba3b1558ca4ff88ba.pdf.
———. 2025. Devising a Strategic Approach to Artificial Intelligence: A Handbook for Policy Makers. Washington, DC: International Bank for Reconstruction and Development. https://documents1.worldbank.org/ curated/en/099060525125542871/pdf/P506884a1130fff-9c6f-4a78-8216-191b979d44b9.pdf.
World Economic Forum. 2025. “Why digital public infrastructure is key to building a connected future.” April 17. www.weforum.org/stories/2025/04/digitalpublic-infrastructure-building-connected-future/.
• Deepak Maheshwari
Zhang, Laney. 2023. “China: Provisions on Deep Synthesis Technology Enter into Effect.” Library of Congress, April 26. www.loc.gov/item/globallegal-monitor/2023-04-25/china-provisions-ondeep-synthesis-technology-enter-into-effect/.
Zugravu, Andreea, Ben Safran, Tarek Mansour, Ary da Cunha and Aurélie Espérandieu. 2024. “Using AI in economic development: Challenges and opportunities.” McKinsey & Company, May 22. www.mckinsey.com/industries/ public-sector/our-insights/using-ai-in-economicdevelopment-challenges-and-opportunities.
About CIGI
The Centre for International Governance Innovation (CIGI) is an independent, non-partisan think tank whose peer-reviewed research and trusted analysis influence policy makers to innovate. Our global network of multidisciplinary researchers and strategic partnerships provide policy solutions for the digital era with one goal: to improve people’s lives everywhere. Headquartered in Waterloo, Canada, CIGI has received support from the Government of Canada, the Government of Ontario and founder Jim Balsillie.
À propos du CIGI
Le Centre pour l’innovation dans la gouvernance internationale (CIGI) est un groupe de réflexion indépendant et non partisan dont les recherches évaluées par des pairs et les analyses fiables incitent les décideurs à innover. Grâce à son réseau mondial de chercheurs pluridisciplinaires et de partenariats stratégiques, le CIGI offre des solutions politiques adaptées à l’ère numérique dans le seul but d’améliorer la vie des gens du monde entier. Le CIGI, dont le siège se trouve à Waterloo, au Canada, bénéficie du soutien du gouvernement du Canada, du gouvernement de l’Ontario et de son fondateur, Jim Balsillie.
Credits
Senior Fellow S. Yash Kalash
Director, Programs Dianna English
Senior Program Manager Ifeoluwa Olorunnipa
Publications Editor Lynn Schellenberg
Graphic Designer Sepideh Shomali
Copyright © 2026 by the Centre for International Governance Innovation
The opinions expressed in this publication are those of the author and do not necessarily reflect the views of the Centre for International Governance Innovation or its Board of Directors.
For publications enquiries, please contact publications@cigionline.org.
The text of this work is licensed under CC BY 4.0. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
For reuse or distribution, please include this copyright notice. This work may contain content (including but not limited to graphics, charts and photographs) used or reproduced under licence or with permission from third parties.
Permission to reproduce this content must be obtained from third parties directly. Centre for International Governance Innovation and CIGI are registered trademarks.
67 Erb Street West Waterloo, ON, Canada N2L 6C2 www.cigionline.org