UNDP’s Human Development Report 2025 asked the world a critical question: Are we on the brink of an artificial intelligence (AI)-powered renaissance, or are we sleepwalking into a future shaped by inequality and diminished freedoms?

At UNDP we've been grappling with this question not just in theory, but in practice. Across the globe, development practitioners are seeing AI move from future promise to present-day tool, one that can go down two very different paths, either increasing inequalities or helping to solve humanity's most pressing challenges.

In partnership with the Global Fund, UNDP has deployed AI-enabled portable X-ray machines for more efficient and accurate tuberculosis screening in remote areas while building digital capacities of frontline workers in Tajikistan, Turkmenistan, and the Pacific region.

In over 60 countries, UNDP has supported AI for rapid assessment of policy alignment between global and national biodiversity targets. This allows governments to reduce time spent on research and to focus on strategic dialogues and stakeholder engagement.

With eMonitor+ deployments in more than 25 countries, UNDP’s AI-powered tool is supporting national and regional partners in analyzing and addressing hate speech, disinformation, and prevalence of  tech-facilitated gender-based violence.  

More broadly, AI is playing an increasingly important role in crisis prevention, response and recovery. In Ukraine, Myanmar, and Haiti, AI tools are employed to predict displacement, assess damage, and analyze debris. Through the Anticipatory Data Hub, real-time analytics are combined with conversational AI, enabling 7,500 practitioners to query risks, track threats, and access insights seamlessly via Application Programming Interface (API)-enabled platform. 

To harness responsible AI use while protecting people and the planet, the UNDP Global Policy Network examined over 50 AI applications developed and supported by teams across the organization. With experience from diverse country contexts, three key lessons emerged:

1. Multidisciplinary teams are essential 

Successful AI requires diverse teams comprising policy experts, AI technology specialists and ‘translators’ who understand both technology and policy. A multidisciplinary, multi-stakeholder approach is key to finding impactful AI solutions. 

Commodity crops such as coffee and cocoa are among the main causes of global deforestation, and growing demand for sustainable production has made traceability a critical requirement. Yet smallholder farmers often lack the tools to participate. To bridge this gap, UNDP piloted the use of AI-generated data layers for coffee traceability in Ecuador, Colombia and Costa Rica to support compliance with the EU Deforestation-free Regulation. The project, built on initiatives like ProAmazonia, brought together cooperative members, government officials, forest experts and supply chain specialists and digital experts. The result? Not the most technologically cutting-edge solution, but one that actually solves a pressing development challenge. 

2. Government ownership and local capacity for sustainability 

For AI solutions to be sustainable, national or local ownership and IT capabilities are essential. 

The Data in Climate Resilient Agriculture (DiCRA) platform exemplifies this approach. This Digital Public Good (DPG) provides open source, open access to geospatial datasets and algorithms for climate-resilient agriculture, helping policymakers and researchers make evidence-based decisions for climate-resilient farming. Supported by over 100 volunteer data scientists and 14 organizations and initially led by the UNDP Accelerator Lab in India in partnership with the Government of Telangana, the platform is now hosted and maintained by India’s National Bank for Agriculture and Rural Development, ensuring sustainability. 

3. Invest in AI governance 

To ensure AI benefits everyone, we must invest in ethical development and governance, while working with countries to build capacity for safe, secure, environmentally sustainable and trustworthy AI systems.

Many institutions struggle with technical expertise gaps and limited engagement in national and global digital policy processes. To address these challenges, UNDP and partners, including the Inter-Parliamentary Union and the Commonwealth Parliamentary Association, are supporting national institutions with training, peer networks, practical oversight tools, and policy guidance to govern AI effectively.

The environmental cost of AI also demands urgent attention. From the energy and freshwater needed to cool data centres to the material demands of hardware, AI carries significant resource implications. The International Energy Agency forecasts global energy demand to double to 945 terawatt hours by 2030. Initiatives such as the Coalition for Sustainable AI are advancing good practices and collective action to reduce their impacts.

AI governance is not just a technical issue. It’s a foundation for inclusive development, accountability, and the protection of human rights in a digital age. 

A call to action

Renaissance or sleepwalk? Our future rests on the choices we make today.

Through the AI Sprint initiative, UNDP is already working with countries worldwide to build AI capabilities. From developing national strategies to training government officials, building data infrastructure, and advising on responsible AI, we're not just deploying technology, we're building ecosystems that will determine whether AI reduces or amplifies global inequality.

At the Hamburg Sustainability Conference in June, global leaders endorsed the Hamburg Declaration on Responsible AI for the SDGs, with UNDP as co-architect. This declaration represents our collective commitment to ensure AI serves sustainable development.

But commitments need action. We call for broader coalitions and partnerships to collectively shape AI to ensure its benefits are inclusive and equitable for all, and that it works for both people and the planet.

The time to engage, collaborate and act is now. 

(With contributions from Reina Otsuka, Osama Aljaber, Manish Pant, Guro Wiik, Minako Manome, Andreas Pawelke, Nina Grinman and Jeremy Boy.)