Locations
AI for Development: UNDP Vision
By Robert Opp | Chief Digital Officer, UNDP
We must bend the arc of AI development to ensure that AI can be harnessed for the benefit of all and does not exacerbate existing divides, reinforcing biases and concentrating technological power in the hands of a few.
Opportunities and Challenges
Artificial Intelligence (AI) has significant potential to enhance governance, improve public services and drive climate action to accelerate progress towards achieving the SDGs. In the past decade, many applications of AI have emerged, particularly in essential areas such as agriculture, healthcare and climate adaptation. For example, Crop Mate, a solution developed in Trinidad and Tobago and recognized through the UNDP Accelerator Lab’s Green Innovation Challenge, provides farmers real-time AI-powered information about soil conditions and automatically recommends nutritional interventions to ensure crop health.
Yet, the rapid expansion of AI also introduces new challenges and exacerbates existing ones, including deepening digital divides, reinforcing biases and concentrating technological power in the hands of a few. The equity gap in AI is staggering. If we consider the case of AI compute, only 2 percent1 of the world's data centres are in Africa and 5 percent of AI innovators in Africa2 have the compute power they require. When considering the linguistic diversity of existing AI models, many are either unavailable or underperforming for the 1.2 billion native speakers of low-resource languages.3
UNDP Global Vision for AI and Sustainable Development
UNDP is committed to bending the current arc of AI development towards equity and overcoming the disparities and access issues in AI. For AI to serve as a force for good, its benefits must reach those who need it the most. Building on its digital portfolio spanning 120 countries, UNDP AI efforts are focused on the following areas:
Data: strengthening data governance and increasing representation of local languages in AI model design and development. Through its Local Language Accelerator program, UNDP is focused on digitizing data in local languages through locally beneficial licensing models.
Ecosystem capacity and readiness: providing capacity building and technical advisory for policy development and creating locally sustainable AI ecosystems.
‘Green’ Compute: expanding access to sustainably designed and deployed AI infrastructure, including access to green energy sources for compute and data centres to address the growing footprint of AI.
Trust and safety: assisting countries to build the foundations for equitable and trustworthy AI. UNDP has launched a global program that advocates for context specific and locally relevant safeguards that prioritize local vulnerabilities.4
Partnerships and financing: building multistakeholder partnerships and directing investments to start-ups and organizations, especially those in the Global South.
To help deliver on the above, UNDP’s AI Landscape Assessments5 provide actionable insights to support countries to address critical gaps in infrastructure, human capital, governance and capacity to help shape their national AI strategies. Through the AI Hub for Sustainable Development,6 an initiative co-designed with the Italian G7 presidency, we are strengthening AI foundations for economic growth in Africa.
Pakistan—The AI Opportunity
Pakistan has made significant strides in digital connectivity over the past decade. According to UNDP Pakistan National Human Development Report (NHDR) 2024, Doing Digital for Development, at the start of 2024, the country had 130 million broadband users, with a mobile broadband penetration rate of 43 percent. The NHDR produced the Digital Development Index7 that shows a strong correlation between districts that perform better in digital transformation and those that have higher human development outcomes. A study by Better than Cash Alliance8 demonstrates that full adoption of digital payments could boost Pakistan's GDP by up to 7 percent. This reflects the country's expanding digital economy.
The AI market in Pakistan is projected to grow by 27.6 percent between 2025-2030, resulting in a market volume of approximately US$3 billion by 2030.9 The Presidential Initiative for Artificial Intelligence and Computing10 aiming to train 100,000 students in AI, blockchain and cloud computing and the draft National AI policy11 demonstrates the country’s clear intention to employ AI for economic growth.
However, overcoming gaps in infrastructure and AI compute needs and continuing to train and upskill the workforce will be key for AI development. For instance, the largest data centre in Pakistan has a capacity of 3 megawatts (MW)12 and the data-centre growth rate of 4.4 percent from the total current base capacity of 20 MW13 is insufficient to meet growing AI demands. Further, funding and support are also needed to boost AI research and local AI development in the country. As per a report from the State Bank of Pakistan in 2023, only 10 percent of new IT graduates are employable in the market.14 In addition, the country also needs to manage workforce transition accompanied by AI-led automation.
The Path Forward
A collaborative approach with multiple stakeholders is essential for AI to truly drive development in Pakistan. Governments, private sector, civil society and international organizations must collaborate to build innovative and inclusive AI ecosystems. Our UNDP Digital Office is committed to working closely with Pakistan.
UNDP is applying AI based machine learning to provide the government of Pakistan with more precise data, deeper insights, in real-time on national development policies. The findings have revealed Pakistan’s SDG progress and have been referenced in the Integrated SDG Insights Report, a national document outlining the country’s SDG status. UNDP Pakistan NHDR 2024, used insights from an AI supported social risk monitor to capture real-time data informing the Report’s analysis.
UNDP Pakistan has also launched a project to localize AI-driven data solutions for multi-stakeholder development planning and budgeting. The project aims to develop an integrative, generative AI platform to support government with real-time data for effective planning, budgeting, monitoring and reporting of SDGs progress across the country.
We stand at a critical inflection point where we must ensure that we bend the arc of AI towards greater equity, both globally and at the national level. The challenge ahead is not just technological, it is about governance, access and capacity. AI’s benefits will not automatically trickle down, instead they must be consciously directed toward reducing inequality and supporting sustainable growth.
1. Africa Data Centres Association, ‘Data Centres in Africa: Focus Report 2024’, Africa Data Centres Association, 2024, [online] Available at: https://africadca.org/en/data-centres-in-africa-focus-report-2024
2. United Nations Development Programme, ‘Only Five Percent of Africa’s AI Talent Has the Compute Power It Needs’, United Nations Development Programme, 2024, [online] Available at: https://www.undp.org/digital/blog/only-five-percent-africas-ai-talent-has-compute-power-it-needs
3. Kocmi, T., Tezcan, A., & Federmann, C., ‘Findings of the WMT 2023 Shared Task on Quality Estimation’, Proceedings of the Eighth Conference on Machine Translation (WMT 2023), Association for Computational Linguistics, 2023, [online] Available at: https://aclanthology.org/2023.wmt-1.40/
4. United Nations Development Programme, ‘AI, Trust, and Safety’, United Nations Development Programme, 2024, [online] Available at: https://www.undp.org/digital/ai-trust-and-safety
5. United Nations Development Programme, ‘AILA’, United Nations Development Programme, 2024, [online] Available at: https://www.undp.org/digital/aila
6. AI Hub for Development, AI Hub for Development, 2024, [online] Available at: https://www.aihubfordevelopment.org
7. United Nations Development Programme, Pakistan National Human Development Report 2024, United Nations Development Programme, 2024, [online] Available at: https://undppknhdr2024.com
8. Better Than Cash Alliance, Realizing the Promise of Responsible Digital Payments for Merchants in Pakistan, Better Than Cash Alliance, 2024, [online] Available at: https://www.betterthancash.org/alliance-reports/realizing-the-promise-of-responsible-digital-payments-for-merchants-in-pakistan
9. Statista, ‘Artificial Intelligence - Pakistan’, Statista, 2025, [online] Available at: https://www.statista.com/outlook/tmo/artificial-intelligence/pakistan
10. Presidential Initiative for Artificial Intelligence and Computing, Presidential Initiative for Artificial Intelligence and Computing, 2024, [online] Available at: https://www.piaic.org
11. Ministry of Information Technology and Telecommunication, National AI Policy Consultation Draft V1, Government of Pakistan, 2024, [online] Available at: https://moitt.gov.pk/SiteImage/Misc/files/National%20AI%20Policy%20Consultation%20Draft%20V1.pdf
12. Baxtel, ‘Jazz Opens New Tier III Data Center in Pakistan’, Baxtel, 2024, [online] Available at: https://baxtel.com/news/jazz-opens-new-tier-iii-data-center-in-pakistan
13. Data Center Map, ‘Data Centers in Pakistan’, Data Center Map, 2025, [online] Available at: https://www.datacentermap.com/pakistan/
14. The Express Tribune, ‘Only 10% of IT Graduates Employable: SBP’, The Express Tribune, 2024, [online] Available at: https://tribune.com.pk/story/2417643/only-10-it-graduates-employable-sbp