Artificial intelligence (AI) could increase global trade by up to 40 per cent by the year 2040, according to the World Trade Organization’s newly released World Trade Report 2025.

However, this potential can only be fully realised if countries bridge existing digital divides, invest in infrastructure and skills, and maintain open and predictable trade policies.

The WTO’s flagship annual publication, which was launched on September 17 during the WTO Public Forum in Geneva, set out a detailed examination of the transformative power of AI and the risks associated with unequal access to it.

In her foreword to the report, WTO Director-General Ngozi Okonjo-Iweala highlighted both the promise and the peril that AI presents. She stated that AI could substantially lower trade costs and enhance productivity, yet the current distribution of access to AI technologies is uneven.

‘With the right mix of trade, investment and complementary policies, AI can create new growth opportunities in all economies. With the right frameworks, trade can play a central role in making AI work for all. The WTO is committed to supporting this effort,’ DG Okonjo-Iweala said.

The WTO’s report forecasted that AI could increase global trade volumes by 34 to 37 per cent by 2040, with especially large impacts on digitally deliverable services, including AI-powered services themselves, which were expected to grow by 42 per cent.

It explained that this increase would stem from reduced operational trade costs, improved productivity, and the expanding market for AI services.

Alongside this trade expansion, the report projected a rise in global GDP by 12 to 13 per cent as AI adoption became more widespread.

However, it warned that these gains were likely to be unevenly distributed unless significant efforts were made to improve infrastructure and regulatory capacity in lower-income countries.

Simulations carried out by WTO economists indicated that policy choices could lead to stark differences in outcomes.

In scenarios where low- and middle-income economies were able to halve their digital infrastructure gaps with wealthier nations and increase AI adoption, income gains could reach up to 15 per cent for low-income economies and 14 per cent for middle-income ones.

In contrast, if no progress were made in bridging these digital divides, income in low-income countries would rise by only 8 per cent, compared to a 14 per cent increase in high-income economies.

The report identified trade as a key enabler of AI diffusion, noting that economies heavily relied on international markets for AI-enabling goods such as semiconductors, high-performance computing equipment, raw materials, and cloud services.

In 2023, global trade in these critical inputs had amounted to $2.3 trillion. It emphasized that access to these goods was essential for AI development and deployment, yet remained largely concentrated in wealthier countries with stronger infrastructure, greater investment, and more specialized expertise.

The WTO further noted that AI was improving trade efficiency and accessibility, especially for small and medium-sized enterprises (SMEs).

AI tools were said to be enhancing supply chain visibility, automating customs procedures, reducing language barriers, and helping navigate complex regulations.

According to a 2025 joint WTO–ICC survey, nearly 90 per cent of firms using AI reported measurable improvements in trade-related activities. More than half stated that AI had enhanced their ability to manage trade risks.

Empirical studies cited in the report suggested that AI might contribute up to 0.68 percentage points annually to growth in total factor productivity.

This increase in efficiency was expected to significantly support economic growth over time, particularly in services sectors such as customer support, software development, and business consulting—areas where AI had already demonstrated strong results.

However, the report cautioned that without proper safeguards, AI adoption could exacerbate global inequalities.

It warned that the benefits of AI might remain concentrated in high-income countries and among large firms in digitally connected urban centres unless targeted policy interventions were made.

The report pointed out that while over 60 per cent of large firms reported using AI, only 41 per cent of small firms did, and in low- and lower-middle-income countries, fewer than one-third of businesses had adopted AI at all.

Labour markets were also expected to be affected by AI. While overall demand for labour could rise—particularly for low-skilled workers in some cases—the wage premium for medium- and high-skilled workers was projected to decline modestly due to task automation.

The report noted that real wages across all skill levels were still expected to rise, but the relative demand for higher-skilled workers might fall as AI took over many of their functions.

This could lead to a 3 to 4 per cent decline in the global skill premium, potentially narrowing income inequality within some countries—though the actual outcome would depend heavily on complementary labour policies and education systems.

The WTO’s analysis found that countries open to digital services trade typically experienced stronger innovation spillovers, as reflected in higher rates of cross-border AI patent citations.

Specifically, it stated that a 10 per cent increase in digitally deliverable services trade was associated with a 2.6 per cent rise in international patent references, illustrating how trade supported knowledge transfer in AI.

Despite these benefits, the report expressed concern over a growing trend of protectionism in AI-related sectors.

It highlighted that quantitative restrictions on AI-enabling goods had increased from 130 measures in 2012 to nearly 500 by 2024, primarily driven by high- and upper-middle-income countries.

Additionally, it noted that some low-income countries still imposed bound tariffs of up to 45 per cent on essential AI-related imports, thereby limiting their access to critical digital tools.

To assess trade barriers more comprehensively, the WTO introduced a new measure—the AI Trade Policy Openness Index (AI-TPOI)—which evaluated restrictions on AI services, enabling goods, and cross-border data flows.

The findings showed large disparities in openness, with middle-income countries often maintaining the most restrictive policies.

Low-income countries appeared more open, though this was attributed more to underdeveloped regulatory frameworks than to intentional trade liberalization.

The report also raised concerns about the concentration of AI capabilities in a small number of countries and firms.

Since 2010, over 98 per cent of government subsidies supporting AI development had come from high-income and upper-middle-income nations.

Likewise, access to critical infrastructure such as data centres and renewable energy sources remained heavily concentrated.

It was noted that data centres already consumed 1.5 per cent of global electricity, and that most energy policies supporting this growth were being implemented in wealthier countries.

Educational systems in many parts of the world were reported to be lagging behind the needs of an AI-driven economy.

The report stated that fewer than one-third of developing countries had adopted national strategies to integrate AI into their education systems.

This skills gap, it warned, could entrench existing economic inequalities. It also pointed out that most competition and intellectual property policies related to AI were concentrated in high-income countries, where AI-specific frameworks were far more likely to have been established.

Despite these challenges, the WTO expressed confidence in its ability to promote inclusive AI development through trade.

The report noted that existing WTO agreements already facilitated the movement of goods and services essential to AI.

It mentioned the Information Technology Agreement, which helped reduce hardware costs; the Technical Barriers to Trade Agreement, which promoted regulatory transparency and harmonization; the General Agreement on Trade in Services, which supported open trade in services; and the TRIPS Agreement, which fostered IP protection and innovation transfer.

The WTO also highlighted its efforts to promote inclusive trade through capacity-building and technical assistance programmes.

Initiatives such as Digital Trade for Africa and the Women Exporters in the Digital Economy Fund were cited as examples aimed at helping developing economies, SMEs, and underrepresented groups benefit from digital transformation.

These initiatives were said to increasingly include AI components, such as in sustainable agriculture and smart logistics.

Nevertheless, the report acknowledged that global cooperation on AI remained limited.

It observed that while regional trade agreements had started to include AI-related provisions, they were generally narrow in scope and geographically concentrated.

The report called for broader multilateral cooperation on trade-related AI policy and argued that updating WTO rules, expanding participation in the Information Technology Agreement, and strengthening commitments under the General Agreement on Trade in Services could help make AI more accessible and affordable worldwide.

The report concluded with a clear warning: the future impact of AI would depend heavily on current policy choices.

It stressed that if inclusive strategies were adopted—ranging from infrastructure investments and workforce development to regulatory alignment and open markets—AI could serve as a driver of shared prosperity and trade-led growth.

However, it cautioned that if the present trajectory persisted, AI could deepen global inequalities and concentrate its benefits in a limited number of economies.