By Dhruv Banerjee - South Asia Desk

Regional Cooperation in South Asia: A Tale of Ups and Downs

South Asia’s struggle to achieve effective regional integration has been well documented. Previously, World Bank estimates have indicated that intra-regional trade accounts for a meager 5 percent of the total trade in South Asia. For context, the percentage for intra-regional trade in South-east Asia is about five times as much, a significant difference. In turn, a combined lack of regional integration and stability has led to a dearth of foreign direct investment flowing into the region. A multitude of reasons have held back regional integration. These include border instability between countries, diverging economic goals, and uncertainty about each other’s intentions

Initially, the path to overcoming these issues was institutionally conceived in the form of the South Asian Association for Regional Cooperation (SAARC), established in 1985. However, since 2014, there have been a total of zero summits, primarily because of increased tensions between India and Pakistan following the 2016 Pulwama terror attack. SAARC’s institutional deadlock was a deeply worrying event with regard to the prospect of regional cooperation in South Asia. SAARC’s failure led to a re-evaluation of neighbourhood policy for countries like India, with the Bay of Bengal Initiative for Multi-sectoral Technical and Economic Cooperation (BIMSTEC) now viewed as an alternative institutional approach. The idea was that connectivity along the Bay of Bengal would be a strategically effective measure to spur regional interconnectivity.  

Over the last nine years or so, India, Sri Lanka, Nepal, Bangladesh, Thailand, Bhutan, and Myanmar have ramped up participation within the organisation. They have focused on several different areas of cooperation, held frequent meetings, and streamlined the organisational structure. There have been six BIMSTEC summits, the latest one taking place in Thailand in April 2025. A notable exception to collaboration under BIMSTEC, however, is the lack of measures around Artificial Intelligence. This is even more surprising given that technology is one of the key areas of collaboration in the organisation. Thus far, there are no concrete initiatives under BIMSTEC that seek to explicitly address Artificial Intelligence as a means of cooperation. There are no expert groups explicitly seeking to address AI, and no official frameworks, agreements, or MoUs specifically aimed towards AI. Not capitalising on the prospect of strengthening cooperation by focusing on AI would be a big opportunity foregone for BIMSTEC. 

Given the rapid diffusion of the technology, most member states do not yet have comprehensive regulatory frameworks in place. This is actually an opportunity, since they can now coordinate their respective domestic policies, which in turn would reduce the compliance costs for companies in a BIMSTEC state to work across borders. Reports on the Asia-Pacific demonstrate that intergovernmental collaboration on AI can have a two-fold effect of attracting foreign investment and improving public service delivery. Converging AI policies can also improve technical and technological transfer between countries. Certainly, global AI summits and the spread of AI Safety Institutes (AISI) seem to suggest a global consensus that AI collaboration is indeed fruitful. All of this is to suggest something intuitive: It is time for South Asia, and more specifically, BIMSTEC, to focus on collaboration in Artificial Intelligence.

How AI Collaboration Can Be Effective for BIMSTEC

Given the breadth and sheer scale of developments related to AI globally, it can be tempting to try to dabble in every possible policy angle related to AI. However, for a multilateral organisation such as BIMSTEC, with relatively limited funding, it would be prudent to focus on a few specific factors. There are four main considerations with regard to AI cooperation in BIMSTEC. 

Firstly, cooperation on AI could lead to regional economic growth and attract more investment from abroad. This is a sentiment echoed by the BIMSTEC Secretary General, who stated that “regional collaboration on AI standards” was essential in achieving “collective development goals”. Certainly, the conditions are conducive for such growth. After all, the BIMSTEC countries have a young population, and therefore a strong workforce, with a majority of people below the age of 30. Collaboration on upskilling this workforce to be adept at developing and using AI is thus a lucrative opportunity for members. A good way to do this would, of course, be to facilitate cross-border exchange programs and education opportunities specifically designed for AI and related fields, perhaps through a flagship BIMSTEC AI scholars program. Providing a platform to AI companies in BIMSTEC countries in the form of a workshop or a startup summit would further allow cross-order exchange of ideas and facilitate a young and vibrant regional AI innovation ecosystem. 

Secondly, collaborating on AI could help BIMSTEC members collectively address issues in critical sectors. For context, BIMSTEC has 7 priority “sectors” of cooperation, which have further sub-sectors. Coordination between members on the use of AI technologies in key sectors such as agriculture, connectivity, and disaster management (sub-sector under security) would be highly beneficial. The 2nd BIMSTEC ministerial meeting on Agriculture called for coordination in using “new agriculture technology, including artificial intelligence”, and such efforts across each of the sectors would make the Bay of Bengal region a hub for incorporating emerging technologies into critical sectors. This, in turn, would also prove attractive to investors. 

Thirdly, AI safety and cooperation are relatively non-political subjects, making cooperation easier. Naturally, issues related to security or even economic integration tend to be harder to build consensus around, due to the potentially politically divergent objectives of members in these areas. A number of security-related conventions and agreements in BIMSTEC have taken a long time to finalise, or have seen really slow progress. In contrast, AI is a relatively less contentious issue. Agreeing on the broad tenets of AI safety, or technology transfer of low-risk Artificial Intelligence, is unlikely to elicit much opposition. Since all the countries within BIMSTEC are developing nations, they also largely share the dilemma of balancing AI for economic growth with regulation to ensure AI safety. 

Photo by Amar Preciado: https://www.pexels.com/photo/detailed-vintage-globe-showing-south-asia-region-30792665/

Finally, AI-oriented collaboration is logistically plausible if the existing measures under the technology sub-sector can be leveraged effectively. Currently, technology is a sub-sector under the broader science, technology, and innovation sector, which is led by Sri Lanka.  Under this sub-sector, the Technology Transfer Facility (TTF) can be particularly useful in the context of AI collaboration. The mandate of the TTF (defined in an MoA signed in 2022) should be modified to include Artificial Intelligence. Following this, it can run “diagnostic studies” to assess the AI needs of each member state. Then, this information can be used by the TTF to provide targeted recommendations for AI-related technology transfer between member states. Additionally, the BIMSTEC Expert Group on Science, Technology and Innovation (STI) Cooperation, which has previously met in 2023 and 2024, could focus on broader objectives. The expert group should gear its third meeting towards addressing the potential of establishing a BIMSTEC framework on AI safety. This would help harmonise the currently divergent AI regulatory approaches undertaken by member states. In this vein, a prudent area to focus on would be to recommend that all the member states set up AI Safety Institutes to become part of the global AISI network, with India having decided to set up an AISI earlier this year. 

Keeping all of this in mind, it is fair to conclude that collaboration on AI can serve as an effective path to increased regional cooperation in South Asia. In particular, BIMSTEC should consider increasing its focus on AI, in terms of both discourse and funding. After all, coordinating approaches to AI would be beneficial for member states in terms of both economic and security considerations, while also providing a relatively non-political area of cooperation. The future can be a bright one for South Asia, provided it works together to embrace the technologies that define it.