• Southeast Asian countries are increasingly looking to AI to boost jobs for lower and middle income populations.
• All 10 ASEAN member states have national AI strategies, increasing competition and potential tension in the AI race.
• ASEAN collectively has a "light touch" governance and ethics framework, as it navigates competing influences from both east and west.

Southeast Asia's emerging economies are vying to become a top AI hub — a race that has them both coming together and, quietly, battling among themselves.

The Association of Southeast Asian Nations (ASEAN), made up of 10 countries with a combined population of 672 million people, already has some advantages when compared to Europe or the U.S.

With over 200 million people aged 15 to 34, the region's youthful and largely tech savvy populations make the region adaptable to future technological advances. That, combined with government support for accelerating AI in the region, could deliver substantial rewards for local workers.

"AI can significantly improve productivity across industries, and this boost in efficiency can lead to an uplift of incomes for all workers," Jun Le Koay from the consultancy Access Partnership and the author of research paper "Advantage Southeast Asia: Emerging AI Leader" told CNBC.

"Additionally, as industries increasingly adopt AI technologies, new jobs are emerging that will require AI skills. This evolution creates opportunities for low-income populations to acquire new skills and transition to better paying positions," he said.

Le Koay added that the AI boom presents opportunities for Southeast Asia to make use of its existing infrastructure. Koay believes that ASEAN countries have made "massive strides" in significantly increasing internet access over the last decade which has "created a digitally native population ready to adopt and innovate with AI."

With smartphone adoption ranging from 65% to 90% in ASEAN nations, AI adoption is expected to rapidly take shape.

Grace Yuehan Wang, CEO at Network Media Consulting and a scholar at the London School for Economics, doesn't foresee any of the ASEAN countries leading the AI race anytime soon.

"ASEAN as a region has demonstrated a strong GDP growth rate in recent years and undoubtedly is one of the most economically prosperous blocs in the world in the foreseeable future," she told CNBC.

A developed digital infrastructure, education of "high-level technical talents in technology industry including AI, as well as world-level universities (both STEM — science, technology, economics and math — and comprehensive universities), successful industrial and research collaborations are some elements still lacking in ASEAN's AI ecosystem," she said.

AI competition between the ASEAN countries is "largely in attracting foreign investment and collaborations with world-leading universities," Wang added.

Singapore steals a march

Ten countries — Brunei, Cambodia, Indonesia, Laos, Malaysia, Myanmar, the Philippines, Singapore, Thailand and Vietnam — make up the ASEAN club. All 10 have published national AI strategies.

Singapore was among the first to unveil its vision, in 2019. The island state updated its plans in December 2023. Ambitions include expanding its AI workforce to 15,000 — triple the current amount — as well as creating research and development centers.

AI Centre of Excellence for the Manufacturing Sector opened in September, with the aim of integrating artificial intelligence across supply chains.

Singapore's AI mission has state backing, with the government promising to invest 1 billion Singapore dollars ($741 million) over the next five years.

The country appears to have stolen a march, "thanks to its R&D, economy, education system, international business standing," said Wang.

Singapore topped Salesforce's 2023 Asia Pacific AI Readiness Index, which evaluated 12 nations. Other ASEAN member states — Malaysia, Indonesia, Vietnam, The Philippines and Thailand — were lower on the list, in the eighth to twelfth spots.

Localized AI for developing nations

Singapore's muscle doesn't appear to have stopped the aspirations of its near neighbors.

Vietnam is hinging its bets on developments in AI, playing to its strength in assembly, testing, and packaging capacities as it meets global demand for chips. The country's national strategy includes ambitions to develop into ASEAN's center for research and development of AI solutions by 2030. The country, for example, has already attracted a $1 billion investment from South Korean manufacturing extending to 2025.

In 2023, VinAI, a part of multisector conglomerate Vingroup, unveiled an open-source language model designed specifically for Vietnamese users called PhoGPT.

The localized alternative to ChatGPT suggests that "English-dominant AI models cannot be applied to all social and cultural contexts, while on a deeper level, it demonstrates the efforts in overcoming fears of widening existing divides and inequalities among the less powerful technology regions and countries," said Wang.

German artificial intelligence translation startup DeepL is already tapping into the regions "rich linguistic diversity," which Chief Revenue Officer David Parry-Jones says is an "asset, fostering a wealth of cultural exchange and deepening regional identity."

Parry-Jones told CNBC that the European startup wants to offer ASEAN AI language models that could boost manufacturing, translating legal documents or support multilingual customer service centers in the region.

"We know companies and governments are looking for best-in-class context-sensitive translation tools so they can continue to grow quickly without being trapped by language barriers," he said.

Other developing countries have their eye on harnessing AI for traditionally labor intensive industries.

For example, Cambodia's 60-page report details how the developing nation wants to harness AI for "social good" and agricultural technology, boosting the sector which represented 22% of Cambodia's GDP and employed about 3 million people in 2018.

Developing ASEAN countries that aren't as digitally-developed as Singapore have greater challenges to becoming AI ready, let alone achieving full blown AI policy.

"There are several regulatory building blocks that need to be intact and robust before credibly embarking on AI," Kristina Fong, lead researcher for economic affairs at ASEAN Studies Centre of ISEAS-Yusof Ishak Institute, told CNBC.

She added that "adverse effects of AI to users could come fast and hard without any institutional oversight" with conversation at state level needed to "effectively manage these rapid developments with minimal societal harm."

Diverging from Europe

ASEAN countries collectively released a regional guide to AI governance and ethics in February. A year earlier, European Union officials on tour in Southeast Asia tried to persuade them to follow the EU's AI regulations.

Instead of being swayed, ASEAN countries argued the EU had been too quick to adopt regulation without fully understanding AI risks.

The Asian bloc has diverged from Europe on AI governance, applying a "light-touch approach seems to be the most apt for the region," said Fong.

"This is mainly due to several factors including the absence of a central legislative body in ASEAN, unlike the EU, as well as the notable differences in digital capabilities and regulatory capacities among the ASEAN member states," she said, adding that Southeast Asia's approach framework on AI ethics "serves more as a practical guide," rather than strict policy.

Wang said ASEAN's AI ethics divergence is not necessarily a battle between picking a Western or Chinese approach. International cooperation, she said, is at the heart of ASEAN's AI ethics framework.

The fundamental challenge ASEAN countries face is "not a technological one, but a political one," said Wang, with the Covid-19 pandemic pushing countries to work together more closely on mutual trade and diplomacy.

What will keep them on the right track to achieving AI plans is retaining their young, savvy population.

Perhaps, Wang argued, a national education strategy to complement AI plans could be most effective.