UNDP: Asia-Pacific's AI surge could widen inequality—millions of jobs at risk even as ~$1T gains beckon

A new UNDP assessment dated 2 December 2025 warns about an uneven AI surge. The report centers on Asia-Pacific, but its message travels. Concentrated gains, unequal capabilities, and fast-moving adoption can widen existing divides. Morocco should study these signals and act early. UNDP sees large upside but warns of unequal distribution. AI could add nearly $1 trillion to Asia's economy over the next decade. It could lift annual GDP growth by around two percentage points and raise productivity by up to five percent in sectors such as health and finance. But benefits today concentrate in heavy investors like China, Singapore, and South Korea. Chief Economist Philip Schellekens is direct. Countries investing in skills, compute, and sound governance "will benefit," while others risk falling "far behind." Kanni Wignaraja reinforces the pace gap. "AI is racing ahead, and many countries are still at the starting line." The report highlights who is most exposed. Women and young adults face significant disruption. Hard-won gains in health, education, and income could slip if adoption beats inclusion. Job displacement and data exclusion add pressure. Asia-Pacific's numbers underscore the stakes. The region hosts over half of the world's population and, by UNDP's account, more than half of all AI users. China holds nearly 70% of global AI patents. Six countries there host more than 3,100 newly funded AI firms. UNDP also flags structural constraints that amplify risk. These include limited infrastructure and connectivity, scarce digital skills, inadequate computing power, and weaker governance capacity. The environmental footprint of AI adds indirect pressures on electricity and water. The agency urges ethics and inclusion before scaling deployments. What does this mean for Morocco? The risk logic applies. AI rewards scale, skills, compute, and governance. Countries that build these pillars will capture more value and cushion disruption. Morocco's digital foundations are improving. Connectivity is widespread in cities and expanding in rural areas. Data protection is established under national law and enforced by the CNDP. Universities and engineering schools are training data practitioners and researchers. An emerging startup scene is testing applied AI. Local teams use computer vision for agriculture and resource monitoring. Others apply machine learning to logistics, retail demand, and credit scoring. These efforts show practical pathways. Agriculture stands out for near-term impact. Satellite imagery and sensor data can optimize irrigation and fertilization. Yield forecasting helps farmers manage inputs and cash flow. Early disease detection protects crops when climate stress intensifies. Public services can also benefit. Voice-based assistants can guide frontline health workers with triage protocols. Simple tools can support vaccination tracking in low-connectivity areas. Document summarization can accelerate case processing in courts and administrations. Tourism and mobility offer quick wins. Multilingual chatbots can improve traveler support across French, Arabic, Darija, and Tamazight. Routing algorithms can reduce congestion and fleet fuel use. Computer vision can enhance safety at transport hubs. Finance and compliance are ready for AI. Fraud detection models can flag unusual transactions while protecting privacy. Risk scoring can help lenders reach underserved small businesses. Explainability and audit trails should be embedded from day one. The UNDP advises simple, resilient tools in lower-resource contexts. Morocco can prioritize offline-capable, voice-first helpers for health and agriculture. These should work with intermittent internet and low-end smartphones. Data plans should remain affordable. A language-first approach is essential. Tools must support Moroccan Arabic (Darija), Tamazight, and Modern Standard Arabic. Clear instructions and localized content matter. Voice interfaces reduce barriers for users with limited literacy. Women and youth need tailored pathways. Training should match local demand and time constraints. Stipends and childcare support can boost participation. Apprenticeships with startups and public agencies can open doors. Jobs will change. Roles in call centers, back-office operations, and routine processing face automation exposure. Transition plans are critical. Internal mobility, short upskilling bursts, and wage support can reduce shock. Build skills with blended programs. Micro-credentials in data literacy, prompt engineering, model evaluation, and AI product management can scale quickly. Universities can partner with companies for applied projects. Public bootcamps can target priority sectors. Compute matters as much as skills. Shared AI labs in universities can provide model training access and hands-on practice. Local data centers can host compliant workloads close to users. Cloud credits and procurement frameworks can lower costs for startups. Data quality drives outcomes. Open, privacy-preserving datasets in agriculture, logistics, and public services can unlock innovation. Transparent licensing and governance reduce risk. Community annotation programs can build local language resources responsibly. Morocco's energy profile can be an advantage. Renewable generation can power data centers and shared compute. Efficient cooling and heat reuse can cut the environmental footprint. Scheduling compute to align with green supply can reduce emissions. Governance must be practical and measured. The CNDP already anchors data protection. Algorithmic transparency, fairness assessments, and risk tiering can extend oversight into AI. Citizens need clear redress channels for automated decisions. Procurement can steer the market. Simple AI risk checks in tenders can screen vendors for safety and inclusion. Requirements can include offline capability, local language support, and accessibility. Outcome-based contracts can keep focus on real impact. Financing should blend public and private tools. Seed grants can de-risk early pilots in health and agriculture. Public procurement can scale effective solutions. International partners can bring cloud credits, training, and research exchanges. Partnerships magnify capacity. Universities can host shared datasets and evaluation benches. Startup accelerators can link teams to customers and mentors. Diaspora networks can help with advanced model work and language resources. A 12–36 month action plan can keep progress grounded. - Publish a practical AI inclusion checklist for public buyers. - Fund five voice-first pilots in health and agriculture with offline support. - Launch micro-credentials for frontline workers and administrators. - Set up shared compute access in two university labs with renewable power. - Release privacy-preserving datasets in agriculture and mobility. - Create grants for women and youth apprenticeships at AI startups. Measure results, then scale what works. Keep metrics tied to service quality, not demos. Track time saved, error reduction, and user satisfaction. Update rules as patterns emerge. Morocco's path does not require frontier models on day one. It needs useful tools, inclusive training, and steady governance. Focus on simple solutions that meet people where they are. Build capacity while protecting vulnerable workers. UNDP's message is clear. Pair growth ambitions with inclusion and ethics. Invest in human capital, affordable compute, and strong governance. This reduces risk while enabling real productivity gains. The region-level warning should be a catalyst. Morocco can avoid unequal outcomes by designing for access and trust. It can build value with applied AI in sectors that matter. Practical steps today shape a safer future. Key takeaways - AI rewards skills, compute, and governance; Morocco should invest in all three. - Start with voice-first, offline-capable tools for health and agriculture. - Focus on local languages and accessible interfaces to widen adoption. - Protect women and youth with tailored training and job transition support. - Use renewable energy and efficient cooling to manage AI's resource footprint. - Embed transparency, fairness checks, and redress channels in public deployments.