How Asean can unlock promise of artificial intelligence

ARTIFICIAL intelligence (AI) is still in its early stages in South-east Asia: the use of one common type of AI - machine learning, where machines are fed data to learn from - is just beginning to have an impact on the region.

While all member states of the Association of Southeast Asian Nations (Asean) engage in some level of AI research and adoption, Singapore is the region's clear leader. Its AI work across multiple industries led Startup Genome's 2017 Global Startup Ecosystem Report to rank Singapore ahead of tech hot spots like Austin, Texas, and Stockholm. However, there is a wide range of digital maturity within Asean.

If AI development were left purely to market forces, early adopters in financial services and in high tech and telecom would be likely to extend their use of the technology. But greater value in other sectors, such as retail and healthcare, would remain largely untapped. We already see AI radically altering the way online retailers fulfill orders, how manufacturers make parts and assemble them, and the way doctors decide how to battle some diseases. And change will only accelerate as the drivers of AI growth - shrinking computing costs and growing data bases - improve.

In South-east Asia, companies in the most digitally savvy industries after high-tech - banking and telecom - have already launched initiatives, but have faced difficulty in scaling up, often because their workforces lack critical skills in data science and business translation or they lack sufficient data.

South-east Asia's financial-services firms have thus far adopted AI primarily to improve the customer experience. Hong Leong Bank of Malaysia, for example, uses AI to detect callers' emotions by the way they speak on the phone. CompareAsiaGroup, a Hong Kong-based startup, uses machine learning to choose the most economical financial, telecom and utility services for clients in five Asean member countries. For AI to have a broad, long-term effect on the sector, South-east Asian banks could integrate AI deeper into their organisations, into functions such as credit scoring, dynamic pricing and digital marketing.

Few Asean banks have scaled up these types of applications. Taking advantage of this opportunity will need established banks to develop new skills and fintech startups - there are already 300 in the region - to innovate.

AI can enable manufacturers to manage their factories remotely, make real-time decisions and improve efficiency. However, this will rely on real-time data from machines equipped with sensors and linked digitally - the so-called Internet of Things (IoT).

Some large regional manufacturers are taking steps to join the IoT, but many others are hesitant. Because labour costs are low, firms do not always see a business case for the capital investment required. This may change as the region develops and populations age, shrinking the labour pool. We already see this in China, where manufacturing wages have doubled in the last 10 years. In response, Chinese firms are expected to invest US$59 billion in robotics by 2020.

PRIVACY LAWS

McKinsey Global Institute research concluded that expanding data use in healthcare could yield more than US$300 billion in value annually around the world. At the same time, AI can improve care, especially in large countries with relatively few doctors and specialists. For example, Indonesia, a country of 250 million people, had only 41 radiation oncologists in 2014.

Despite the potential benefits, common use of AI in patient care remains years away in Asean. The region lacks enough consolidated data to support the use of advanced analytics, let alone AI. Hospitals have data, but usually on paper, and they have little incentive to share it. Most Asean countries require data to stay "onshore", hobbling efforts to build region-wide data sets that could improve AI analysis. Sharing data, even in anonymised form, also could run afoul of social norms or local privacy laws.

Governments have a role to play, too. Well-crafted strategies and policies can turn today's innovation into sustainable growth for Asean over the longer term, and AI can provide a major boost to productivity. Governments can facilitate this by setting aspirational national and regional goals and spurring private-sector innovation and adoption.

Rapid urbanisation in Asia, for example, exacts a high price in productivity lost to congestion - from 2 to 5 per cent of GDP. Cities are responding with AI systems to manage their buses, subways and traffic. Private-sector companies are helping. Ford, for one, is pivoting from being a carmaker to being a "mobility" provider. Its City Solutions division seeks to use AI to integrate mobility options from public transit to ride hailing to bike sharing.

As for businesses, adopting AI is not a simple proposition. Companies will need to make fundamental changes including embracing data- driven decision-making, inculcating a culture with a "test and learn" approach, striking innovative partnerships with the government or specialist firms to incubate the scarce skillsets needed for AI efforts, and prioritising effort and investments to strengthen the data infrastructure. No single organisation acting alone can solve all of the issues surrounding these technologies.