AI Singapore: How a small island plans to bridge AI research and reality

• AI Singapore’s programmes will help researchers, industry and the public sector work together to better leverage the technology’s potential

• Defining problem statements in key areas will help to focus AI research in Singapore

• As AI becomes more prevalent, society will need to address the social, moral and legal issues associated with its adoption

What if you could optimise every aspect of your life? Imagine a system that could automate personal purchases, making adjustments based on usage and price fluctuations, or that could monitor and execute financial investments without you having to lift a finger. Imagine if small business owners, CEOs and government officials could make complex decisions simply by asking a computer to analyse cross-correlations in multiple datasets.

With the availability of big data and the use of artificial intelligence (AI) to gain insights from it, these Hollywood-worthy scenarios are quickly becoming reality. Recognising the potential for these disruptive technologies to address a wide range of societal and industrial challenges, Singapore in May 2017 launched AI Singapore (formerly AI.SG), the country’s maiden venture into harnessing the capabilities of AI.

Speaking at the Big Data and AI Leaders Summit on 31 October 2017, Mr Koo Sengmeng, Deputy Director (Strategic Alliances) at AI Singapore, laid out the city-state’s strategy for using these new digital tools to improve the lives of residents and optimise business processes.

Bridging research and reality

Mr Koo noted that AI talent is scarce globally, not to mention in a relatively small country like Singapore. This being the case, partnerships between experts from industry and academia are especially pertinent, and efforts in the AI space need to be kept focused and relevant in order to maximise their potential, he added.

“One of the primary objectives of AI Singapore is to pull companies and researchers together to perform deep, inspired research, eventually improving business practices,” said Mr Koo.

At the same time, the initial investment into AI ventures can be costly and difficult for many companies. To help them overcome this barrier to entry, AI Singapore has started 100 Experiments (100E), a programme that matches companies in need of AI expertise with AI researchers.

“100E conducts short-term collaborative efforts geared towards proving the business value of AI. In this case the problem that we want to solve is not a problem of technology, but of business. We are actually taking on the entire financial risk of doing the AI research for you,” said Mr Koo, adding that businesses will thus be better equipped to leverage Singapore’s financial and intellectual capital. This does not mean that there is no contribution from the companies that wish to embark on this venture. In order to retain the intellectual property, and as an indication of responsibility and interest, there is a minimum financial investment required. Mr Koo, however, encouraged the sharing—through open-source databases, for example—of any ideas or developments that arise from such collaborations, as this would be in the nation’s best interests.

Rising to grand challenges

To guide research efforts, AI Singapore is also working with the public to define a series of Grand Challenges—important issues in areas such as healthcare, urban mobility and finance that can be effectively addressed using AI.

“What we want to achieve through the Grand Challenges is to uncover technical as well as financial surprises that will help us focus the questions in AI research, and allow us to break through any fundamental bottlenecks,” said Mr Koo, adding that the research should also have a tangible outcome—a product or application—that can benefit society.

Yet, Mr Koo acknowledged that the process is not as straightforward as it sounds. “It is in itself a Grand Challenge to find Grand Challenge questions!” he said, while adding that the rewards will be well worth the effort. “If we ask the right questions, we will uncover the right problem statements.”

More than just megabytes

As AI becomes more widely adopted, society will need to address the social, moral and legal issues associated with the technology, said Mr Koo. “If we are going to involve AI in decision making, we need to know how we are going to regulate AI-driven business or healthcare decisions.”

To illustrate the complex scenarios that can arise, Mr Koo gave the popular example of autonomous vehicles reacting to an imminent accident. “Does the AI make the decision to kill the driver and save the pedestrian, or the other way round? These will be interesting questions that we need to deal with.”

Given these complexities, Mr Koo emphasised the need for planning ahead and for government agencies to be involved in identifying such tricky situations and their ramifications. He also warned that if AI is to be an effective and holistic solution, it cannot be studied or applied in a silo. “We cannot merely look at AI research—we also have to involve sociologists, economists and other disciplines,” he concluded.