**Introduction** Governments can prevent artificial intelligence (AI)-driven digital inequality by strengthening the enabling conditions that determine who can adopt, deploy, and benefit from AI. The most effective policy responses focus on digital infrastructure, skills and organizational capability, data governance, market structure, and the ability to participate in evolving digital trade and regulatory frameworks. Without coordinated action across these areas, AI adoption risks reinforcing existing gaps between large and small firms, advanced and developing economies, and connected and underserved regions[1][2]. **Policy levers to prevent AI-driven digital inequality** **1.** **Strengthen digital and computing foundations** Unequal access to digital infrastructure and computing capacity remains a primary source of AI-related inequality. Reliable broadband, resilient data infrastructure, and affordable access to cloud and computing services are prerequisites for AI adoption across firms and public institutions. Where access to these inputs is limited or costly, AI use is concentrated among large firms and advanced economies[1][2]. Governments can reduce these barriers through shared public procurement of cloud and computing services, support for regional data centers, and targeted access schemes for universities, startups, and small and medium-sized enterprises. Investment in digital public infrastructure, including interoperable digital identity and data platforms, further supports inclusive participation in AI-enabled services and markets[3]. **2.** **Expand skills and organizational capability** AI-driven inequality is closely linked to shortages in digital skills and firm-level capability. Workforce strategies that combine basic digital literacy, technical training, and continuous skills upgrading are essential to diffuse AI adoption beyond frontier firms. These efforts are most effective when aligned with labor market demand and delivered through vocational systems, higher education, and employer co-investment[2][4]. Small and medium-sized enterprises face additional constraints due to limited technical and managerial capacity. Targeted advisory services, applied training, and technology extension programs can support practical AI adoption in operations, compliance, and export activity. Strengthening public-sector capability in data management, procurement, and AI oversight also helps ensure that AI use in public services does not exclude groups with lower digital access or literacy[4][5]. **3.** **Use data governance and competition policy to widen participation** Access to data and market structure are central determinants of AI adoption and diffusion. Clear and predictable data governance frameworks that permit lawful data use and sharing lower entry barriers and compliance costs, particularly for domestic firms and smaller providers, while fragmented or uncertain rules raise fixed costs and reinforce concentration[1][6]. These concentration pressures are amplified by scale economies and data advantages in AI-related markets, which tend to favor incumbent firms. Competition policy and pro-competitive regulation in digital and cloud markets can limit lock-in and exclusion, while public procurement can support wider participation by emphasizing interoperability, open standards, and contract structures accessible to smaller suppliers[5][7]. **4.** **Reduce exclusion risks through international rule engagement** AI adoption is increasingly shaped by international standards, digital trade rules, and regulatory cooperation affecting data flows, digital services, and compliance requirements. Limited capacity to engage in these processes can translate into higher compliance costs and reduced market access, particularly for developing economies and smaller firms[1][6]. Governments can reduce these risks by investing in regulatory and negotiating capacity, prioritizing interoperability over fragmented requirements, and linking new digital and AI-related obligations to technical assistance and capacity-building mechanisms. These steps help prevent AI-related rules from becoming de facto barriers to participation in digital trade[6][8]. **Conclusion** Preventing digital inequality from AI requires a coordinated policy approach centered on digital and computing infrastructure, broad-based skills and adoption capability, inclusive data governance, competitive market structures, and effective participation in international rule-making.
