A $ 5 billion AI industry can save SL economy

A strong market oriented Government could make the AI industry a $10 billion one within five years 

“I believe whatever smart, ambitious people are working on will be the trend of the future. I do think that it's worth thinking critically about what the future will be.”

-Sam Altman

On the same day that my previous article titled "Sri Lanka must immediately transcend the IMF Agenda" was published in this newspaper, the Central Bank Governor Dr. Nandalal Weerasinghe held a press conference. It was purely coincidental that both Dr. Weerasinghe and I expressed similar sentiments albeit in different terms. 

Dr. Weerasinghe, in the same press conference, extended an invitation to individuals who believe that taxes are exorbitant to propose alternative solutions. I have already touched upon this matter in some of my prior articles in the Daily FT, and now I intend to delve into it with greater specificity.

This article discusses the prospect of establishing a thriving $ 5 billion industry locally, specifically catering to the global market's demand for Artificial Intelligence (AI) solutions. Although challenging, this endeavour is far from impossible - even at this late stage. It necessitates two fundamental components: unwavering determination and a comprehensive national-level strategy. While this article does not present a fully-fledged proposal, it highlights key considerations that Sri Lanka should seriously contemplate in its pursuit of finding solutions to the ongoing economic crisis.  

Why AI?

This was not the first time Sri Lanka had faced a national level economic crisis. It’s a déjà-vu. The things were as bad as or perhaps worse in the early 1880s. All that has been earned in the fair-weather times have gone down the drains. Foreign currency reserves have dipped to the bottom. Unemployment has skyrocketed. The country was learning the danger of depending on a single commercial crop, the hard way. An unfamiliar fungal disease Hemeleia Vastatrix (later known as ‘Devastating Emily’) has almost completely destroyed the coffee plantations. The crisis was so grave that of 1,700 British planters, all except 400 returned to England empty-handed. 

How did the authorities react? They gave up on coffee and experimented with other possibilities like cocoa, cinchona and tea. The last one emerged the champion. The country progressively transformed from a coffee-based economy to a tea-based one, within just a decade. The economy was booming to a level that new pathways in railways were supported by the revenue of the plantation industry.

AI in 2024 shows the same promise shown by tea nearly a one and a half century ago. Predicting the exact size of the global AI industry in 2024 is like gazing into a crystal ball, but based on current trends and expert predictions, it's expected to be significantly larger than what it was in 2023. Applying a conservative 20% growth rate to the lower end of the 2023 market size (around $ 400 billion) would yield a market size of approximately $ 480 billion in 2024. However, considering higher growth estimates and other factors, it could easily reach $ 500 billion or even surpass that.

If Sri Lanka were to grab at least one hundredth of the global AI market, that would amount to an additional export revenue of $ 5 billion, coming closer to what is earned by the textiles and garment sector at present. The potential is surely higher than this and depends on how much focus we would give to that sector. With determination and careful policy planning, a strong market oriented Government could make the industry a $10 billion one within a period of five years. No other imaginable industry in Sri Lanka demonstrates a massive potential of this kind. 

What are Sri Lanka’s strengths and opportunities in AI?

Sri Lanka’s software industry dates back to the early 1980s and has become a significant player in the global market, attracting both local and international companies. Sri Lanka's tech ecosystem thrives on a blend of talented professionals, somewhat supportive of governmental policies, and a burgeoning start-up culture. These may not be sufficient, but are the necessary conditions at the starting point of a vibrant AI industry. 

Says Export Development Board (EDB) website: The Sri Lankan ICT sector serves a number of industry verticals with over 500 companies at present. These include communication, apparel and textiles, banking, financial services and insurance (BFSI), healthcare, manufacturing, media, retailing, transportation, travel and leisure and many more. Sri Lanka's success in ICT exports are chartered through an impressive upward trend in annual figures. The export revenue of this sector grew from $ 166 million in 2006 to reach $ 1.2 billion in 2021 with a workforce of over 120,000. With over 90% value addition and high paying jobs, it has made a significant impact on the growth of the Sri Lankan economy. 

Sri Lankan ICT companies export software products and services to regions such as North America, EU, Australia, East Asia, the Middle East, Africa and the Nordic region. It says further, with their core competencies covering automated application testing, infrastructure outsourcing, high-end R & D, enterprise resource planning (ERP), cloud technology, mobile applications and many other business solutions, Sri Lanka also acts as an offshore development centre for several fortune 500 companies from the USA, Ireland, UK, Australasia, Sweden and joint venture development companies from Sweden, Norway, USA, Japan, etc.  

One of Sri Lanka’s key opportunities is its proximity to India. Whether we like it or not, in the near future, India is projected to emerge as a leading developer and consumer of AI. Establishing and maintaining robust trade connections with India would prove invaluable. Notably, Sri Lanka's capital, Colombo is a mere one-hour flight from Bangalore, India's prominent ICT hub. This proximity surpasses the distances to India's other major metropolitan areas, such as New Delhi, Mumbai, and Kolkata. It would be remiss not to recognise and earnestly consider the potential of this opportunity.

Can a country cradle a new $ billion plus industry?

Consider any successful industry in an Asian country, and it is highly likely that its growth was initially fostered by government intervention rather than occurring naturally. Applying this perspective to industries in Sri Lanka, we must reflect on the historical periods of British and Dutch colonial rule. In the post-colonial era, Sri Lanka has not been proactive in embracing and developing many industries, with the exception of a notable case such as the apparel industry in the late 1980s.

President Premadasa, despite lacking prior experience in leading industries, demonstrated his visionary leadership by spearheading the growth of a new industry from its inception. Notably, he emphasised the establishment of factories in rural areas, thereby promoting regional development. President Premadasa personally oversaw the resolution of infrastructure challenges, ensuring that garment factories received priority in power supply from the CEB. Additionally, new roads were constructed, and local banks provided loans to support the industry's growth.

Perhaps a better example from South Korea

 Another international example comes from South Korea. It was the Japanese who introduced shipbuilding to Korea during its colonial period in the 1910s. Still, post-independence Korea has never been a key shipbuilder until it has been identified as a ‘priority industry’ under the third five year plan (1972-76). 

The original idea of building Korean supertankers arose in one of the many conversations between Hyundai Chairman Chung Ju Yung and then Korean President Park Chung-hee, sometime in the early 1970s. The anecdote goes; Park gave the other only one option if the other did not want the corruption charges against the firm thoroughly investigated. Get into shipbuilding. Chung toured the world looking at shipbuilding and searching without success for financing for a Korean yard, until a deal was reached with a Greek ship-owner in April 1972 to build two 250,000-ton tankers. Production on these ships began in March 1973.  

The boom in shipbuilding in Korea was dramatic. In 1972, Korea manufactured small cargo ships of only 21,000 tons of gross weight. Two years later, it built ships of over 600,000 tons in total weight. It peaked at 800,000 tons in 1975 before falling back to a little over 400,000 tons a year, for the rest of the years until President Park’s death in 1979. By 1976, however, Korea had the capacity to produce 2.6 million tons of ships a year. Now, Korea is the second largest shipbuilder, following China. The industry has been recognised as one of the most competitive in the world, with the overall competitiveness of the Korean shipbuilding industry among the highest, particularly in R&D, design, and production.

So the fact that Sri Lanka has never done it before must not prevent us from attempting. It is precisely our unwillingness to take significant risks that has resulted in our current status as a middle-income nation. In order to progress from a developing economy to a developed one, we must disrupt this cycle and take a risk for a change.

Why AI to be treated a PRIORITY INDUSTRY?

I write PRIORITY INDUSTRY in capitals to provide it the prominence it truly deserves. Let’s get back to the Korean case. South Korea’s turning point was the Heavy and Chemical Industries Drive (HCI Drive) launched under the third five-year plan (1972-76). For this they identified a few industries (iron and steel, shipbuilding, petrochemicals, machinery and equipment etc.) and treated them at a different level. Development of these selected industries became the work of the Government and its officers. More than 60% of the Korean President’s working time was invested in resolving issues for these industries. They were provided loans at substantially lower interest rates. Necessary policies were introduced with the minimal delay. It was this special treatment that turned them into giants.

Mainstream economists generally favour a different approach than the one above, as it contradicts the theories they learned in College. They tend to support free markets with minimal Government intervention and strongly oppose Government discrimination. According to their perspective, the market itself should determine which industries are viable, allowing others to naturally phase out. This viewpoint is based on the straightforward logic that dinosaurs became extinct while mammals survived. However, it is important to note there is limited evidence to support the effectiveness of this theory in Asia. With the exception of Singapore, most developed Asian nations have achieved their level of development by practising the opposite of what traditional textbooks suggest. Korea may be an extreme example of preferential treatment of industries and protectionism, while other Asian countries have adopted similar approaches to varying degrees.

Allocating a significant portion of resources to AI without a clear strategic plan may not yield substantial benefits. Merely providing funding alone would not foster the development of the AI industry in Sri Lanka. It is crucial to establish AI as a priority sector, comparable to how Korea prioritised and supported the shipbuilding industry in the 1970s. 

What level of priority must AI industry receive?

We talk about developing a $ 5 billion industry from almost scratch, within a span of three years.  A priority industry benefits from strong Government backing, that helps overcome obstacles and challenges. Currently, the AI industry per se does not receive any specific incentives or guaranteed infrastructure support. It is treated like any other industry, say for example, apparel. An example highlighting the lack of prioritisation is the software industry's request for a dedicated fuel station during a fuel crisis two years ago. This request, which would have ensured a reliable fuel supply for generators and vehicles, was not granted, despite the software industry being considered a priority sector. A genuine priority classification entails providing the necessary support and protection to industries that are susceptible to political, social, and economic shocks. By doing so, these industries can focus on their operations and growth while Government officials address other pressing issues. It is crucial to recognise the importance of enabling a conducive business environment for priority industries and ensuring that bureaucratic procedures do not hinder their progress.

Further, by recognising AI as a priority industry, Sri Lanka can implement targeted policies and initiatives that would effectively drive the growth and advancement of the AI sector in the country. This includes fostering a supportive ecosystem, investing in research and development, promoting skill development, and facilitating collaboration between the academia, industry, and Government agencies.

Do we have HR to back a $ 5 billion AI industry succeed?

The simple and straightforward answer is a NO. Not yet. There is a substantial gap between what it requires and what we have for building a $ 5 billion industry in a relatively small market in Sri Lanka, both in terms of technical and HR. Still, that understanding is a good starting point. We can plan what kind of capacity we talk about. 

Let’s first focus on HR. Estimating the necessary AI workforce for a country like Sri Lanka requires some basic calculations. While there is no specific data available for Sri Lanka, we can look at India as a reference point. India's software industry is substantial, reaching $ 250 billion in 2023, and is supported by approximately 5 million software engineers. Assuming that the value addition from one professional in the AI field is at least twice that of a software engineer, we can estimate that the AI industry would require a trained quality workforce of around 50,000 professionals to reach a value of $ 5 billion within a span of three years.

This goal is not far-fetched for Sri Lanka, considering that the country already produces around 20,000 software professionals annually. With a clear focus on AI and substantial Government backing, both state and non-state universities in Sri Lanka can easily produce the required number of professionals. To achieve this, it would be necessary to implement different programs aimed at reskilling and upskilling trained programmers in AI. These programs would help bridge the gap between existing skills and the specific requirements of the AI industry. It is also important to address the challenges faced by the software industry and mitigate the potential risk of a workforce exodus. 

Given the appropriate emphasis, support, and training initiatives, Sri Lanka has the capacity to cultivate a proficient AI workforce of 50,000 professionals, thus catalysing a $ 5 billion AI industry. However, it is important to acknowledge that the demand for AI expertise may extend beyond this estimate. In such circumstances, it would be prudent to establish a mechanism for attracting professionals from neighbouring countries such as India and Bangladesh through work visas. Sri Lanka should proactively engage in trade pacts with these nations to facilitate this collaborative exchange of skilled professionals.

What about the Infrastructure for a $ 5 billion AI industry?

Probably you have guessed it. The answer here too is a NO. Still, just like in the case of HR, Sri Lanka must build most of the necessary infrastructure. While local telecommunication companies are readily poised to deliver 5G services upon sufficient demand and international bandwidth infrastructure is readily available for acquisition, Sri Lanka's data centre landscape perhaps remains a critical bottleneck in realising its full potential as a vibrant AI hub at the level that we talk about. Despite the undeniable excitement surrounding the current state of Sri Lanka's data centre scene and its promising potential for further growth, certain crucial limitations hinder foreign investor confidence. Notably, the issue of power stability and limited bandwidth are routinely emphasised as significant challenges before potential foreign investors take the plunge.

To successfully propel Sri Lanka into the next phase of its AI ambitions, a paradigm shift in data centre design and implementation is imperative. Embracing cutting-edge technologies and practices can address the existing limitations and pave the way for a robust and resilient data centre infrastructure capable of supporting a burgeoning AI ecosystem.

How does this align with the ongoing IMF arrangements?

A critical consideration in evaluating this solution is its compatibility with the existing IMF program. While there is no inherent conflict between the two, this solution does offer potential mitigation for some of the anticipated challenges arising from the program.

It was Milton Friedman who articulated the notion that our current system increasingly burdens the working population with taxes while subsidising those who do not engage in productive work. What strikes me as surprising is the stark contrast between the neo-liberalism championed by the IMF and the neo-liberalism espoused by Friedman himself. The IMF wholeheartedly endorses the transfer of a substantial portion of professionals' earnings into state coffers, even to the extent of slashing 36% of their secondary income derived from sacrificing their leisure time. The rationale behind this approach is rooted in the IMF's belief that excessive inflation is indicative of individuals having an excessive amount of disposable income (a point on which the IMF aligns with Friedman), and the only way to curb inflation is to impose maximum taxation on the populace. However, the economic theories underpinning this approach are subject to debate. Regrettably, these theories do not dissuade the Government from exerting pressure on the people to pay their taxes. Therefore, Dr. Weerasinghe is correct in asserting that anyone who wishes to avoid exorbitant taxes must propose alternative means of generating revenue for the country.

Mainstream economists may not always recognise alternative methods to address an economic crisis beyond taxation and monetary expansion. When in need, rather than resorting to begging, borrowing or stealing, individuals have the opportunity to pursue the more challenging route of earning money. Similarly, countries have the ability to shift their strategic direction and foster economic growth. I hope this piece provides an answer to Dr. Weerasinghe’s question. If we do not want to pay excessive tax, we must prepare the country to earn a $ income.

Conclusion: This is the time to act!

Despite possessing significant potential, Sri Lanka's development trajectory has not translated into the realisation of its full potential as a developed nation. While various factors have contributed to this, let’s not get into that debate. More importantly, it is not inevitable that Sri Lanka remains classified as a middle-income country. A critical turning point is necessary to propel the nation towards a more prosperous future, and achieving this sooner rather than later is of paramount importance. 

The sentiment surrounding the development of AI remains optimistic in Sri Lanka. It is widely acknowledged among almost all political camps, even in the left-wing, that AI is a valuable tool, and the prevailing consensus is that now is the opportune moment to delve into its potential. Therefore, it appears that there are no significant hindrances impeding the implementation of these ideas, aside from the potential risks involved. It is essential to recognise that risks are inherent in such endeavours, but it is equally important to remember that no progress can be achieved without some level of challenge.

(The writer [email protected] is an academic cum a policy researcher. The ideas expressed here are personal.)