Priorities for U.S. Participation in International AI Capacity-Building

While the United States debates engagement in international fora and focuses inward, China is quietly building the infrastructure of global artificial intelligence (AI) influence. In July 2024, China secured the adoption of its resolution on AI capacity-building by the United Nations General Assembly, co-sponsored by more than 140 countries. In April, Chinese President Xi Jinping directed his government to "help Global South countries enhance their technological capabilities" and contribute to bridging the global AI divide. By July, China unveiled the Global AI Governance Action Plan to solidify its place as the world's convener for AI governance. Meanwhile, the U.S. struggles to maintain basic continuity in its international AI engagement, a strategic miscalculation with profound consequences for American influence in the technology that has inevitably set a precedent for breakthroughs defining the 21st century.

AI capacity-building encompasses the development of educational programs and training for local researchers, as well as the provision of compute infrastructure such as data centers and cloud services accessible to local institutions. It also includes support for indigenous AI development by local companies, regulatory and institutional frameworks, and technology transfer that builds local capabilities rather than creating dependencies.

China has operationalized its AI diplomacy through concrete initiatives, including workshops in Shanghai and Beijing that have drawn participants from more than 40 countries, the AI Capacity-Building Action Plan targeting economically developing nations, and the Group of Friends for International Cooperation in AI Capacity-Building, which convenes regular meetings. Russia, despite facing severe technological constraints, has formalized AI cooperation agreements with China and established the BRICS AI Alliance Network, which spans 14 countries. Both nations have strategically leveraged existing infrastructure partnerships, including China's Digital Silk Road and Russia's energy and mining agreements across Africa and Latin America, positioning themselves as viable long-term partners for AI capacity-building with Global Majority countries——that is, nations in Africa, Asia, Latin America, the Caribbean, and Oceania—that collectively represent the majority of the world's population.

China and Russia possess significant potential to expand their global AI influence through educational and cultural exchange mechanisms. Both countries maintain extensive international scholarship programs and cultural initiatives targeting students from the Global Majority, with Russian programs tracing back to Cold War-era soft power strategies. For example, the U.S.S.R. sponsored scholarships for working-class students from newly independent nations in Africa, Asia, and Latin America to attend the Patrice Lumumba Peoples' Friendship University (RUDN) in Moscow, founded in 1960. Given widespread plans by Global Majority countries to enhance education as a key part of AI capacity-building, China's and Russia's established academic programs in AI and related fields could address critical gaps in AI skills development and technical education.

The United States, by contrast, has systematically deconstructed the institutional capacity necessary for sustained international engagement. The U.S. Agency for International Development (USAID), historically the primary vehicle for digital development initiatives, has been shut down, leaving many (if not all) of these programs without funding. The State Department's Global AI Research Agenda, a framework developed to promote international collaboration in research on the societal and economic implications of AI, is non-operational. Key programs such as the Partnership for Global Inclusivity on AI (PGIAI), launched with major tech companies in 2024, operate with unclear status. Projects such as USAID's Responsible AI program, as well as other programs focused on combating misinformation and disinformation, promoting human rights, fostering cultural exchanges, and developing policy frameworks, face uncertain prospects under the current administration's priorities.

The current administration's approach reveals a fundamental gap in understanding what effective capacity-building requires. While the U.S. has focused on electrification, telecommunications infrastructure, and data center investments, it has largely neglected educational upskilling and support for local AI research capacity across Global Majority countries. As a result, African researchers are increasingly utilizing models such as the Chinese-owned DeepSeek due to their more competitive pricing compared to that of American alternatives. U.S. support for the development of computational infrastructure that is accessible to local researchers in resource-constrained settings would enable Global Majority researchers to build their own models. However, given the lengthy timelines required to build data centers, mandates for American AI companies to expand exports of small-parameter, open-source, high-quality models that researchers across Global Majority countries can adapt would provide a faster solution.

Why Global Majority AI Partnerships Matter for U.S. Interests

The United States has compelling strategic reasons to invest in AI capacity-building partnerships with Global Majority countries. These nations collectively represent the majority of the world's population and will comprise the majority of future AI users and developers. The AI governance frameworks, technical standards, and ethical norms established through these partnerships will shape global AI development for decades. Without meaningful U.S. engagement, these frameworks may be shaped primarily by Chinese and covert Russian influence, potentially embedding authoritarian governance models as default approaches to AI development and deployment.

Beyond their influence on governance, these partnerships serve core American security and economic interests. Capable, stable partners with robust AI ecosystems create more reliable allies than nations dependent on adversarial powers for critical technology infrastructure. The growing markets for AI technologies and collaborative research in these regions represent significant economic opportunities that U.S. withdrawal cedes to competitors. Perhaps most fundamentally, global AI governance frameworks will require broad-based participation from countries worldwide and diverse stakeholders across academia, civil society, and industry to achieve legitimacy and effectiveness. A system of AI governance developed among wealthy nations and imposed on the Global Majority will lack the political legitimacy necessary for implementation and will fail to address the diverse contexts in which AI systems operate. Regulatory frameworks designed without input from the countries where they will be applied cannot account for varying legal traditions, institutional capacities, enforcement mechanisms, or societal priorities.

The Strategic Cost of Withdrawal

Rising authoritarianism across Sahelian Africa, Central Asia, and Latin America has been bolstered by China's and Russia's influence through military support, propaganda campaigns, and exports of surveillance technologies. With China already demonstrating its capacity as a leader in promoting international AI cooperation, U.S. withdrawal from multilateral engagement and reduced aid spending heighten opportunities for China to intensify AI partnerships with Global Majority countries. Meanwhile, deepening coordination between Russia and China creates pathways to embed anti-democratic ideologies through the guise of AI capacity-building. For example, this might manifest when AI infrastructure investments come bundled with governance models that concentrate data access and algorithmic control within government ministries, with limited provisions for independent auditing, public transparency, or protection against misuse for political surveillance.

The consequences of the United States’s failure to meaningfully participate in global AI capacity-building efforts extend beyond diplomatic dismantling. China has already announced AI-powered agriculture projects in Kenya and Nigeria, with stated interest in supporting AI development for health care across Africa. During the July BRICS Summit, China and Brazil announced plans to establish a joint AI research facility focused on agricultural development. Global Majority countries are already embracing Chinese leadership in AI. The Nigerian government, for example, has expressed support for Chinese AI governance initiatives. Indonesia has sought Chinese assistance in AI development for aquaculture and agriculture, along with increasing development of AI-enabled smart cities. These efforts will shape how AI systems are developed, deployed, and governed across regions representing the majority of the world's population. And, right now, China is far ahead of the United States in forming these relationships.

Recent U.S. initiatives reveal both the potential and limitations of the country’s approach to forming similar international partnerships. In November 2025, the State Department announced a partnership with California-headquartered Zipline, operating under a pay-for-performance model. This partnership would provide up to $150 million to expand AI-enabled medical supply deliveries across Africa, contingent on African governments signing $400 million in contracts. While this partnership may increase African countries' access to life-saving technologies, garnering nearly half a billion in contracts may be unfeasible given the high debt burden across the continent that limits national spending on essential social services like health care. By contrast, the U.S. International Development Finance Corporation's $90 million equity investment in Cassava Technologies in December 2024, alongside Google and Finnfund, demonstrates an alternate model. This investment supports an African-based company that is now leveraging technology from Nvidia, a U.S. company, to build local capacity. The U.S. government also maintains a $300 million investment in Cassava's subsidiary, Africa Data Centres, which supports data center expansion on the continent. This approach builds local economic capacity, creates employment and technical expertise within African institutions, and advances technological sovereignty by ensuring that critical AI infrastructure is owned and controlled by African entities, demonstrating that U.S. technology transfer can support rather than supplant local capacity building. The current and future administrations should expand approaches like the Cassava deal, prioritizing investments in African-based infrastructure and AI companies with a special focus on those that can integrate American technologies.

The Trump administration's AI Action Plan, released in January 2025 as a comprehensive strategy for maintaining American AI leadership, emphasizes increasing AI exports to allies and partners. However, the emphasis on exports reveals a fundamental misunderstanding of what Global Majority countries actually need. While technology exports can play an important role, they must be integrated with capacity-building efforts rather than serving as standalone solutions. The focus on technology sales without corresponding capacity building reflects a transactional approach that fails to address underlying institutional, regulatory, and human capital gaps, preventing effective AI adoption and governance. As expressed in numerous AI strategy frameworks from Africa, Latin America, and Asia, Global Majority countries require support for developing indigenous research capabilities, training local talent, establishing regulatory frameworks, and building the institutional infrastructure necessary for responsible AI development. As independent capabilities continue to emerge within Global Majority countries, access to American AI products, including support for open-source AI development and data center infrastructure, can support rather than substitute for local capacity building.

A recent executive order following the plan mandates that "American AI technologies, standards, and governance models are adopted worldwide." This mandate, however, is difficult to comply with because the U.S lacks comprehensive federal AI legislation. The government demands global adoption of American standards while simultaneously withdrawing from multilateral mechanisms necessary for collaborative development. This creates an untenable proposition: Countries are expected to embrace American governance models that don't meaningfully exist while navigating visa restrictions, tariffs, and export controls that make genuine partnership nearly impossible.

In May 2025, the Trump administration rescinded its January Diffusion Rule implementing tiered export controls. The original framework disproportionately marginalized Global Majority countries by placing them mostly in Tier 2, which implemented quota-based access to advanced AI chips (like H100s), requiring licenses but allowing deployment within certain caps. Embargoed and sanctioned countries were relegated to Tier 3, resulting in near-total bans on advanced AI hardware and model weights. Despite the recissions, various countries made changes to avoid U.S. economic and immigration restrictions from the Trump administration. For example, Malaysia increased controls on chip smuggling to China, and Eswatini and Rwanda agreed to accept deportees from the United States. China's approach offers a stark contrast that is appealing, rather than threatening, to Global Majority countries. According to Semafor, Microsoft's planned $1 billion data center investment in Kenya has yet to begin construction despite targeting a May 2026 operational date. The article notes that for publicly listed companies such as Microsoft, government contracts alone don't justify investments of this scale, raising concerns regarding the long-term feasibility of the Zipline deal's conditional structure. Meanwhile, China has demonstrated consistent capacity for rapid delivery on large-scale infrastructure projects, including the construction of hydroelectric power plants, shipping ports, railroads, airports, and road networks. As the Semafor reporting observes, China’s long-term geoeconomic interests have trumped concerns around immediate financial returns.

Research on Chinese engagement with African nations further supports these observations. A RAND Corporation analysis of Chinese activities in Africa documents how China's approach emphasizes respect for sovereignty and non-interference in domestic governance, contrasting with Western partnerships that often attach political or economic conditions. This approach, combined with predictable long-term funding commitments, makes Chinese partnerships attractive even when they come with their own complications and dependencies. More broadly, this dynamic reinforces the partnership approaches that many Global Majority countries seek to embrace.

What Authentic AI Partnerships Require

The U.S. has a long history of foreign capacity building across Africa, Asia, Latin America, the Caribbean, and Oceania, and has committed numerous efforts to digital capacity building over the past decade. However, many of these programs now face uncertain futures. The status of the Digital Connectivity and Cybersecurity Partnership remains unclear following the State Department's dismissal of diplomats and experts from the Bureau of Cyberspace and Digital Policy in July 2025, which effectively split apart the bureau responsible for the initiative. The AI Connect program, another initiative run by the State Department with the Atlantic Council as an implementation partner, has not held events since November 2024, leaving its operational status uncertain. The U.S.-ASEAN Science, Technology, and Innovation Cooperation Program remains active but is scheduled to end in 2027. The Promoting American Approaches to ICT Policy and Regulation initiative is no longer active. And the Digital Transformation with Africa initiative, tied to the U.S.-Africa Leaders Summit, faces an uncertain future amid heightened tensions following the Trump administration's decision to disinvite South Africa from the 2026 Group of 20 Summit.

Despite this retreat from international partnerships, Global Majority countries have not abandoned interest in U.S. collaboration. Countries such as Malaysia have expressed interest in working with both the U.S. and China on AI to maximize the benefits that both countries can contribute. This strategy of diversifying partnership options and accelerating progress toward AI capacity-building is likely to be adopted by other Global Majority countries seeking to avoid overdependence on any single partner.

Effective U.S. engagement in international AI partnerships will demand fundamental changes in its current approaches. First, partnerships must prioritize technological sovereignty, investing in capabilities that enable countries to make independent choices about their AI futures. Countries such as India, Kenya, the United Arab Emirates, and Brazil, along with regional bodies such as the African Union, the Association of Southeast Asian Nations, and the Organization of American States, have consistently emphasized aspects of digital sovereignty as a core part of their respective AI frameworks.

AI sovereignty can serve the interests of both partner countries and broader U.S. strategic goals by building independent capacity for AI development. While the concept is context-dependent, African, Caribbean, Latin American, and South Pacific countries—for instance—remain far from the technological ability to produce semiconductors and chips independently. However, a primary objective of AI sovereignty is to reduce reliance on foreign companies while enabling independent control over AI development. U.S. technology transfer can still play a crucial role by providing GPU access, supporting data center construction, and facilitating technology partnerships. But these efforts should focus on building local economic capacity rather than creating permanent dependencies.

While the administration's "American values" remain undefined, meaningful AI partnerships must promote approaches that enhance human rights, privacy, and civil liberties rather than undermine them. The key to achieving these goals while respecting sovereignty is to support locally defined frameworks rather than imposing U.S. models. This means U.S. support should focus on enabling countries to develop their own approaches to transparency and accountability through AI systems that are explainable, auditable, and subject to democratic oversight as defined by local stakeholders.

In practice, this requires funding training programs for local AI auditors, regulators, and oversight personnel who understand both the technical aspects of AI systems and the local legal and cultural contexts. It means supporting multi-stakeholder dialogue that brings together government, civil society, academia, and the private sector to develop locally appropriate standards that inform AI development. The U.S. can also facilitate knowledge exchange between government institutions such as the National Institute of Standards and Technology and Global Majority regulatory bodies, through research collaborations to develop interoperable governance approaches.

Perhaps most critically, the U.S. government must demonstrate commitment to long-term collaborative relationships that transcend political transitions. Global Majority countries need stability to develop robust AI ecosystems, which requires predictable, multi-decade partnerships rather than initiative-driven programs tied to specific administrations. This means avoiding the administrative upheavals that have historically undermined U.S. credibility in international development, establishing institutional continuity mechanisms, and creating funding structures that survive electoral changes.

The Future of U.S.-International AI Engagement

The United States faces a substantial and widening gap in AI capacity-building efforts across Global Majority countries. European national development agencies, including the Swedish International Development Cooperation Agency, the United Kingdom's Foreign, Commonwealth, and Development Office, Germany's Agency for International Cooperation, and Canada's International Development Research Centre, have established comprehensive, multiyear commitments to support AI research centers, access to computing, and digital governance initiatives. The Africa Initiative within the EU's Horizon Europe program starkly demonstrates the disparity in the U.S.’s approach to AI capacity-building: 500.5 million euros have been allocated across 24 calls for proposals to strengthen African-European institutional partnerships through collaborative research. In contrast, American governmental engagement remains fragmented and inadequately funded, with $15 million in funding announced to support AI capacity-building, followed by $33 million in funding announced from the inoperational PGIAI.

Organizations such as Mozilla, the Gates Foundation, and Luminate have invested tens of millions of dollars in AI advocacy, research, civic engagement, and entrepreneurship programs across Africa, Latin America, and South Asia, often filling gaps where U.S. government support is absent. Simultaneously, various UN agencies have launched significant AI governance initiatives across the Caribbean, Africa, and Latin America. The absence of meaningful American governmental participation in these initiatives represents a strategic oversight with long-term implications for U.S. influence in global AI governance discussions. Without consistent governmental engagement, the United States loses the opportunity to contribute to a shared vision for the norms, standards, and regulatory mechanisms that will govern AI deployment, frameworks that require diverse perspectives and collaborative development to achieve legitimacy and effectiveness.

American tech companies have invested substantially in digital skills training, research centers, university partnerships, digital and cloud infrastructure, funding academic research, and supporting AI startups. While these initiatives may align with strategic national interests, corporate initiatives often operate according to business imperatives, like market access and entry. These efforts remain vulnerable to shifts in corporate strategy, trade tensions, regulatory changes, or profitability assessments that could prompt companies to withdraw or redirect resources. This underscores the need for systematized government efforts that can provide institutional stability and long-term commitment.

The current administration's prioritization of impeding Chinese growth over building genuine partnerships is a strategic miscalculation. While concerns about Chinese AI expansion are legitimate, a purely reactive approach focused on containment rather than offering compelling alternatives fails to address why many Global Majority countries find Chinese partnerships attractive: predictable funding, respect for sovereignty, and commitment to long-term engagement regardless of political changes. For the United States to regain meaningful influence in global AI capacity-building, there must be a fundamental shift from viewing international AI engagement primarily through the lens of great power competition to recognizing it as essential for democracy, equitable innovation, and global stability. Without immediate course correction, the United States risks becoming increasingly marginalized in global AI governance discussions and excluded from collaborative partnerships shaping AI development across the majority of the world's population.