A new digital divide grows as AI adoption accelerates
Artificial intelligence is moving from labs into daily life at extraordinary speed. A new United Nations Development Programme assessment warns that the gains, risks and rules are spreading unevenly, and that the divide between countries could widen if the world fails to act. The report, The Next Great Divergence, argues that the pace of adoption is measured in months, not decades. Many countries do not yet have the infrastructure, skills or governance capacity to capture the benefits while cushioning the shocks. That combination, rapid deployment with uneven readiness, is a recipe for deeper gaps in income, opportunity and security.
The warning comes after five decades in which many lower income countries edged closer to wealthier peers through trade, technology and development gains. Those convergence gains are now at risk. The report identifies capability as the key fault line. Countries that can build skills, computing power and credible governance can harness AI to lift productivity, improve public services and spark new industries. Countries lacking those fundamentals risk getting left behind as others set standards, attract investment and reap early dividends.
Asia Pacific sits at the center of this transition. The region is home to more than half the world’s population and more than half of global AI users, yet only about 14 percent of people in the region actively use AI tools. That leaves roughly 3.7 billion people on the sidelines. One in four residents remains offline. In South Asia, women are up to 40 percent less likely than men to own a smartphone. Large social challenges add urgency, including 1.6 billion people who cannot afford a healthy diet and 27 million youths who remain illiterate.
Where the capability gap shows most
Capability means more than bright ideas. It includes affordable and reliable electricity, resilient connectivity, data, cloud and compute access, skilled workers, and institutions that can set and enforce rules. Some economies are sprinting ahead on this front. Others are still building basic digital access and literacy, which limits how quickly they can benefit from AI tools in health, finance, agriculture and education.
Asia Pacific as ground zero
The region’s innovation footprint is expanding fast. China accounts for nearly 70 percent of global AI patents, and more than 3,100 newly funded AI companies have emerged across six economies. Modeling in the report suggests AI could lift annual GDP growth in the region by around 2 percentage points and raise productivity by up to 5 percent in sectors such as health and finance. ASEAN economies alone could capture nearly 1 trillion dollars in additional GDP over the next decade if adoption is broad and inclusive.
Those headline numbers sit alongside stark realities. About 1.3 billion workers in the region are in informal employment with little protection if jobs change or vanish. Nearly 770 million women are out of the labor force, and around 200 million people live in extreme poverty. Industries that rely on routine tasks face pressure from automation. Garment manufacturing hubs, which often employ women, could see faster deployment of robotics. Business process outsourcing, call centers and back office services in countries such as the Philippines, India and Bangladesh face competition from increasingly capable AI systems. Without policies for reskilling, transition support and new job creation, disruption could outweigh near term gains for many workers.
Who is most exposed to AI disruption
The report’s labor analysis points to two groups with particular vulnerabilities, women and young people. Jobs held by women are nearly twice as exposed to automation as those held by men because women are overrepresented in roles with routine, predictable tasks. In several economies, youth employment is already declining in high AI exposure roles, especially among people aged 22 to 25, which threatens early career pathways. Basic digital skills are an added barrier. Only about one in four urban residents, and fewer than one in five rural residents, can complete simple spreadsheet tasks. That skill gap limits the ability to adapt as AI reshapes work.
Philip Schellekens, Chief Economist for UNDP’s Asia and the Pacific Regional Bureau, underscored the risk that a new divide will replace years of progress.
We think that AI is heralding a new era of rising inequality between countries, following years of convergence in the last 50 years.
He warned of broader consequences if the world allows the gap to widen.
If inequality continues to rise, the spillover effects of that in terms of the security agenda, in terms of undocumented forms of migration, will also become more daunting.
Kanni Wignaraja, UN Assistant Secretary General and UNDP’s regional director for Asia and the Pacific, described the pace gap in stark terms.
AI is racing ahead, and many countries are still at the starting line.
The risks are not only about job counts. They also involve who gets seen by the data that guides decisions. Rural and indigenous communities, and many women entrepreneurs, are often underrepresented in the datasets used to train machine learning models. The report cites AI credit scoring models trained mainly on urban male borrowers, which then misclassify women and rural farmers as higher risk. That kind of bias can block access to finance, compounding disadvantage. One in four companies surveyed also expects job losses linked to AI automation, a signal that many employers plan to use AI as a substitute for some tasks.
Benefits are real, but unevenly distributed
AI is already improving government services when deployed with care. Bangkok’s Traffy Fondue platform has handled nearly 600,000 citizen reports, helping city teams fix potholes, clear trash and respond to local problems more quickly. Singapore’s Moments of Life service reduced new-parent paperwork from roughly two hours to about fifteen minutes, a concrete efficiency gain. In Beijing, digital twins, a virtual representation that mirrors a physical system in real time, are helping planners simulate floods and traffic flows to guide infrastructure decisions. These examples show how AI can improve responsiveness, lower administrative burden and deliver measurable public value.
The opportunity extends beyond city management. In health, pattern recognition can accelerate diagnosis of conditions like tuberculosis from medical images. In education, adaptive tools can tutor students in remote schools. In agriculture, AI systems can analyze soil, weather and pest data to advise farmers on planting and fertilizer. Food systems, disaster relief and social protection programs can benefit from real time analytics that help get resources to the right places at the right time.
Yet the distribution of these benefits is uneven. Many countries depend on imported models that do not reflect local languages, dialects or cultural context. That can reduce accuracy and trust, especially in high stakes uses such as health and finance. Local language support requires data and compute, both of which are scarce in many lower income countries. Energy and water demands from rapid data center growth add a further challenge. Scaling cloud and AI infrastructure without straining power grids or stressing water supplies is a delicate balance, particularly where climate goals are already under pressure.
Governance lag and security risks
Rules are catching up slowly. Only a limited number of countries have comprehensive AI regulations in place. By 2027, more than 40 percent of global AI related data breaches may stem from misuse of generative AI, according to the report. Generative systems can create convincing text, images and audio. The same capability that helps a teacher draft lesson plans can also be used by criminals to accelerate phishing attacks, spread deepfake fraud or craft malware. Stronger governance frameworks, including clear accountability, testing standards, auditability and privacy protections, are urgent priorities.
Global participation in setting those rules is uneven. A UN trade and development analysis finds that just 100 companies, mainly in the United States and China, account for about 40 percent of private investment in AI research and development. Meanwhile, 118 countries, mostly in the Global South, are absent from international discussions on AI governance. The global AI market is projected to reach 4.8 trillion dollars by 2033. If a small group of firms and countries define norms and infrastructure while others watch from the sidelines, the outcome will likely magnify existing divides in technology, finance and influence.
The UNDP report sums up the challenge plainly.
The central fault line in the AI era is capability. Countries that invest in skills, computing power and sound governance systems will benefit, others risk being left far behind.
What countries can do now
Closing the gap is possible. Every country has different starting conditions, from electricity access and internet reliability to the strength of schools and public institutions. A practical strategy begins with basics, reliable power, affordable connectivity and devices. It continues with people, learning programs that teach digital skills, problem solving and safe use of AI tools. It includes compute, through local data centers, regional hubs or access to shared resources. It adds inclusive data, so models reflect rural communities, minority languages and women’s entrepreneurship. Finally, it requires credible guardrails that build trust, including transparency, safety evaluations and clear recourse when systems cause harm.
Five priority actions
Governments and partners can move on multiple fronts at once. The following actions are widely applicable, even though implementation should reflect local context.
- Invest in affordable broadband, resilient electricity and community access points so households and small firms can go online reliably.
- Expand compute access with regional AI hubs, public cloud credits and research networks, so universities, startups and public agencies can train and adapt models.
- Scale reskilling and upskilling, with an emphasis on women and youth, pairing foundational digital skills with sector specific training for health, agriculture, education and small business.
- Fix bias, collect and govern data that represents rural, indigenous and low income communities, and require lenders to validate AI credit scoring against discrimination.
- Localize AI, support tools in local languages and dialects, fund open datasets and open source projects that meet public needs.
- Strengthen guardrails, adopt privacy laws, algorithmic transparency, impact assessments and independent audits for high risk uses in public and private sectors.
- Protect people in transition, expand social protection, job matching, wage insurance and incentives for firms to create new roles as automation takes hold.
International cooperation can accelerate progress. UN trade and development experts propose a shared global facility to provide equitable access to computing power and AI tools. That kind of pooled infrastructure could help close the compute gap for researchers, regulators and small enterprises in lower income countries. Public disclosure standards for AI systems, similar to environmental and governance reporting, can also raise transparency and accountability across borders.
Comparisons across regions
Inequality in AI readiness is not only an Asia Pacific issue. Data from Europe show sharper preparedness in countries such as Denmark, Germany and Switzerland, which score well on digital infrastructure, skills and governance. Eastern European countries including Albania and Bosnia and Herzegovina lag behind their western counterparts on many of those indicators. The pattern echoes the capability logic identified by UNDP, where countries with strong starting conditions pull ahead unless others receive targeted support.
There are also lessons in how small states can punch above their weight. Singapore has paired high quality connectivity with public sector innovation and clear guidance for responsible AI. South Korea has combined industrial policy with skills programs and investment in compute. These examples show that policy choices matter. Technology alone does not deliver broad based social gains without complementary investments in people, institutions and fair competition.
Key Points
- UNDP warns unmanaged AI could widen inequality between countries by deepening gaps in capability, skills, compute and governance.
- Asia Pacific hosts more than half of global AI users, yet only about 14 percent of people in the region actively use AI tools.
- Women and youth face higher exposure to automation, and basic digital skills remain a barrier for many workers.
- Bias in training data can misclassify women entrepreneurs and rural farmers, limiting access to finance and services.
- AI can lift growth by around 2 percentage points in Asia Pacific and boost productivity, but benefits are uneven without inclusive policies.
- Examples from Bangkok, Singapore and Beijing show AI’s public service value when deployed with clear objectives and safeguards.
- Only a limited number of countries have comprehensive AI rules, and by 2027 more than 40 percent of AI related data breaches may involve misuse of generative AI.
- Global AI investment and governance influence are concentrated, with 100 companies behind about 40 percent of private R&D funding and 118 countries absent from governance discussions.
- Priority actions include investment in power and connectivity, compute access, reskilling, inclusive data, localization and strong guardrails.
- International cooperation, including shared compute facilities and disclosure standards, can help close capability gaps and spread benefits more evenly.
