India’s international trade in times of AI

It is trite now to say that artificial intelligence (AI) has been the biggest economic story of the past five years. AI is expected to reduce many white-collar jobs, determine how many people will work, what they will work on, and how the services sector will operate. However, apart from the trade in intangibles, humans will continue to trade in physical products and use them in their personal, commercial, and work lives. Merchandise trade will continue to grow as standards of living rise across the world, especially in the developing and least developed countries. The main constraint, and rightly so, on the volume of goods trade, would be environmental concerns. Nevertheless, global merchandise trade flows, too, would be affected by the proliferation of the technology of AI.

Impact on supply chains: In manufacturing and distribution, AI is already making demand forecasting, inventory management, procurement, and supply management more agile. Complex estimations for these activities can be done in times that are orders of magnitude smaller than earlier. If the overall supply chain is to reap the benefits of such efficiency, the logistics of international trade and governance by agencies like Customs and port authorities will also have to be equally efficient. Such government agencies will need to have AI-enabled but human-supervised processes for faster and better decision-making.

As of now, Customs assessment of most of the cargo is facilitated through automated risk assessment. While the details of such systems are confidential, it is reasonable to assume that they have been using machine learning for some time now. Human intervention is limited to the cargo flagged by the system. For such flagged cargo, jurisdiction-agnostic faceless assessment has been put in place. While faceless assessment has been implemented by Indian Customs, feedback from traders indicates that more steps are necessary to reduce inefficiency and avenues for corruption. After the assessment of cargo comes the step of physical examining cargo. This is a sensitive step, as goods must be physically examined not only from the point of view of Customs revenue but also for restrictions and prohibitions related to national security, banned substances, compliance with standards, etc. For this reason, it is essential to distinguish between the various kinds of risks posed by different kinds of cargo or the profiles of importers and exporters. Examination orders generated by the automated system are added on to by human agents (Customs officials) considering relevant facts and circumstances.


A very close coordination between automated systems and human judgment is needed to ensure that validated and secure supply chains are not disrupted due to delays at this stage. To be AI-ready and compete with other countries, continuous upgradation of the automated risk assessment system and its interface with the human (faceless) assessment is needed.

Related parties: Most large value import transactions take place between related parties located across borders. Related parties in simplified Customs and tax language are parties that have common shareholding or direct or indirect mutual control. A long-standing issue that Indian Customs and many Customs administrations the world over have always struggled with is the correct valuation of imports from related parties.

Even in the age of AI, Customs administrations move at a snail’s pace in such matters. As is well known, the corporate income tax authorities would like to assign the lowest possible value to such imports, while Customs authorities would like to assign a value as high as possible to the same transaction. Decision paralysis has plagued Customs’ valuation of related-party imports for years and even decades. Assessments have not been finalised as investigations have continued interminably. In recent years, the Central Board of Indirect Taxes and Customs (CBIC) has made attempts to streamline the process, but these efforts have not yet yielded results. Given the large number of pending cases of related party imports, it may be a good idea for the government to institute a deemed finalisation of assessment if a certain amount of time has elapsed and the importer has provided all required information.

Semiconductor supply chain: Major economies today want to have some part of the semiconductor value chain in their jurisdiction. Part of the reason behind this desire is the critical role AI and its hardware would play in technological sufficiency vs. external dependence and in the security of digital infrastructure. While it is not possible to locate the end-to-end manufacture of semiconductor products in any one jurisdiction, large enterprises are putting up semiconductor fabrication or assembly, test, marking, and packaging facilities in major countries, including India. Notable names include PSMC and Qualcomm, both in partnership with the Tata Group and Micron Technology.

Even as the global value chain of semiconductor chips is getting fragmented, chips as well as specialised machinery like lithography machines, other foundry equipment, wafer dicing and thinning equipment, and testing tools, will continue to be traded across borders. While restrictions by countries that control existing technologies will determine what can or cannot be imported, Indian authorities, for their part, can pave the way for smooth imports into and exports out of the country.

The general trade issues cited above are also relevant to the semiconductor industry in India. Given the criticality of this industry, the time sensitivity of its operations, and the fact that large, reliable companies are involved in this sector, perhaps, in addition to resolution of the aforementioned issues, sectoral schemes of trust-based trade facilitation could be implemented by the Government of India. This is especially relevant as the semiconductor industry had recently escalated its trade and Customs issues to senior levels in the Government of India.

The Age of AI needs trade governance that deploys natural intelligence in conjunction with artificial intelligence for a quantum jump in speed and quality of decision making.

The writer is an independent trade expert.