Promise and potential pitfalls of AI in global trade

Artificial Intelligence is widely regarded as a revolutionary technology that will fundamentally change how we trade across borders. According to the World Trade Organisation (WTO), generative artificial intelligence could contribute an estimated USD 4.4 trillion annually to the global economy, reshaping industries and international trade.

But AI is no magic bullet, and its much discussed powers of disruption may not be universally positive. In fact, it's not out of the question that the widespread adoption of AI will lead to widening trade imbalances between nations and increasing protectionism. At the moment, it's difficult to know which scenario - positive, pessimistic or a mixed bag - will win the day.

The promise of AI

The potential of AI is well known, though adoption is in its early stages (and trailing the hype by some distance). According to Goldman Sachs Research productivity boosts of 25% might be expected.

When productivity rises, economic growth and global trade tend to follow. And there are several other reasons to believe the positive scenario when considering AI's impact on international commerce:

• International trade is partly an exercise in trend spotting, and predicting who will want what, where, when and in what quantities. AI can scan huge datasets to pick out the patterns and trends that can help to accurately predict the hotspots of global trade.

• AI has the potential to simplify supply chain management. AI algorithms can scan a wide range of data to identify the supply chains and delivery routes that are the most optimal in any particular circumstance. This could reduce costs while creating more reliable trading networks.

• AI enhances visibility and control. When combined with Internet of Things (IoT) technology, AI can track goods as they cross countries, continents and oceans. Buyers can follow their orders in real time, helping them prepare for and mitigate against delays or disruption.

• AI can enhance customer-facing communications. Intelligent AI-driven chatbots can talk to customers in their own language, giving them shipping updates and up-to-the-minute delivery estimates and answering any questions they may have.

Widening the technology gap

These are just some of the ways AI can be used to streamline global trade, but there are potential downsides to AI adoption too.

Perhaps most obviously, AI could further widen productivity and trading gaps between the developed and developing world.

"For the foreseeable future, wealthier nations are likely to spend the most on AI," says senior Atradius economist Theo Smid. "If the technology fulfils its promise, the resulting productivity and efficiency improvements will give businesses in those countries a significant competitive advantage over rivals in the developing world."

Left behind nations may struggle to catch up. AI software can be relatively inexpensive to use, but in a world where 2.6 billion people still have no access to the internet, making the most of the technology will require major investments in infrastructure. Many developing nations will struggle to make these investments - which might include electricity generation and data storage capacity, as well as broadband - in the short term.

At the other end of the chain, businesses need access to credit to buy or upgrade their IT equipment. They also need talent. Skilled people implement AI tools and keep them running, and analyse the data they produce. Here, too, investment is required. Like any new technology, in the short term AI is likely to exacerbate a global IT skills gap that already exists.

Narrowing trade networks

In this scenario, left behind nations become less competitive in international markets, limiting global trade growth. AI algorithms that continually scan vast amounts of data (from financial documents to weather patterns) to identify the most cost-effective and reliable producers will quickly weed out any that might be considered a risk.

Algorithms may also be used to identify political or reputational risk. Geopolitical tension is already creating a more polarised trading world, with the US and Europe on one side and China and Russia on the other. AI could make it easier to eliminate supply chain links in 'unfriendly' countries and improve near-shoring and reshoring efforts. Quite simply, erecting trade barriers may be easier in an AI-driven world.

Considerations around data privacy, intellectual property and national security also come into play. The US and other countries have already imposed export controls on advanced chip-making technology to slow the advance of China's semiconductor industry. AI could make international collaboration on trade restrictions more effective. It could even be used to slow its own global proliferation.

IP issues could undermine trade

Challenges around intellectual property (IP) could also lead to greater protectionism. Differing approaches to IP could cause tensions between nations, leading to trade restrictions in digital goods and new barriers on the export of digital services.

Problems might arise from IP in the AI tools themselves, or from goods that straddle physical and digital worlds, like self-driving cars and industrial robots. These real world products use highly sophisticated AI and could in future face import and export bans based on national security or IP protection considerations.

That's not happening just yet, but trade policy in these areas is fragmented, confusing and ripe for modernisation. Even in the absence of new trade restrictions, the inconsistency of current policies around AI and digital data flows may serve to discourage international trade or increase its costs.

AI: impetus or impediment to global trade?

Proponents would argue that, regardless of these challenges, AI's productivity-boosting potential trumps all other considerations. This view is shared by the likes of the Organisation for Economic Co-operation and Development (OECD), which advocates further liberalisation of trade in information and communication technology to make the most of the AI revolution.

But there are clearly obstacles to surmount if this optimism is to be realised. Much of the world could be left behind as AI advances, stunting the potential of global trade growth. On the other hand, AI tools could help developing nations compete, if governments and private enterprise also invest in infrastructure and skills.

In other ways, AI could reinforce a new age of protectionism, by giving countries the means to identify vulnerabilities in their supply chains and create barriers against goods from countries that don't meet new IP or national security standards. Friendly countries could also collaborate more easily and hasten the polarisation of global trade. AI provides the intelligence to make onshoring and near-shoring more straightforward.

With this in mind, the role AI plays in international trade will depend largely on wider geopolitical developments. With a technology that has the potential to both enhance or limit cross-border commerce, politicians and policy makers will ultimately have a major say in what its impact will be.