Singapore unveils National AI Strategy

Launch of Singapore's National AI Strategy by Heng Swee Keat, Singapore’s Deputy Prime Minister and Minister for Finance. Image courtesy of MCI.

Yesterday, at the Singapore Fintech Festival, Deputy Prime Minister Heng Swee Keat unveiled Singapore’s National AI Strategy, which outlines the country’s AI plans as part of the greater Smart Nation initiative. Along with this, the government has set up a National AI Office under the Smart Nation and Digital Government Office, which will be responsible for the strategy’s realization.

The strategy is essentially a 10-year AI plan that aims to make Singapore a global leader in AI by the year 2030. Already, the Singapore government has invested SG$500 million in AI, as part of the Research, Innovation and Enterprise 2020 (RIE2020) plan. And although the strategy, a 45-page document, does not disclose any financials, it gives a very clear indication as to how the initial $500m will be used.

Singapore’s AI strategy will be a combination of both direct government investment into AI, for both public administration and in specific sectors of the economy, as well as growing public-private sector partnerships.

It lists three purposes to the strategy: to implement national-level projects, to foster collaboration between the government industry, and research institutions, as well as to manage new risks that might arise as a result of AI.

Building an AI ecosystem

The strategy aims to build an AI ecosystem in which research, industry and government collaborate in all areas of AI, from building data-sharing partnerships to educating the general public on AI. A large part of this involves industry digital transformation, for which the government hopes develop startups and encourage Small and Medium Enterprises (SMEs) to use AI for their businesses.

One interesting way the government intends to do this is by implementing an initiative called AI Makerspace, a new platform with pre-built solutions and curated data sets that startups and SMEs will be able to use, which AI Singapore is currently developing. This will be complemented by other similar schemes such as Tech Depot on SME Portal, where local SMEs can access “Ready-to-Go” AI solutions developed by A*STAR. These will be implemented in addition to a variety of other programmes such as incubation and coaching.

The 9 sectors and 5 projects

The National AI Strategy states early on, in the Vision, that the government will focus on “key sectors with high social or economic value for Singapore, and for which we can build on existing strengths. sectors of high value and relevance to our citizens and businesses”.

These are: (1) Transport & Logistics, (2) Manufacturing, (3) Finance, (4) Safety & Security, (5) Cybersecurity, (6) Smart Cities & Estates, (7) Healthcare, (8) Education and (9) Government.

Within these sectors, the strategy has identified five major projects that will be implemented, with a timeline consisting of 3 phases: now till 2022, 2022-2025, and 2025.

These 5 projects are: (1) Intelligent freight planning, (2)Seamless & efficient municipal services, (3)Chronic disease prediction & management, (4), Personalized education through adaptive learning & assessmentand (5) Border clearance operations.

1. Intelligent freight planning

In the freight sector, the National AI Strategy aims to pool and dynamically assign trucking jobs, use AI for truck routing and scheduling, and use AI for urban planning. As with all the other projects, this is split up into 3 phases; by 2022, the government hopes to have a common data platform for the logistics ecosystem. By 2025, it hopes to deploy AI applications at the sea gateway, and by 2030, the air and land gateways.

2. Municipal Services

The strategy’s municipal services project will be heavily reliant on IoT, at least at the later stages. Its primary focus will be on estate maintenance, and the project’s first phase will be dedicated to implementing an AI-chatbot for the public to report municipal issues. Later on, in 2025, sensors will be deployed for predictive maintenance of housing estates, with AI to play a role in planning the living environment by 2030.

3. Chronic disease prediction & management

For the healthcare sector, the government is focusing exclusively on chronic diseases, and is for the time being focusing on diabetes and cardiovascular diseases. AI will be used to compute personalized risk scores for chronic diseases, assist in chronic disease management, and provide clinical decision support for primary care doctors.

To that end, the government will, by 2022, deploy its SELENA+ (Singapore Eye Lesion Analyser) application for diabetes retinoplathy screening across the nation. Following that, it will develop a retina-based risk score application for 3H-related (high glucose, high blood pressure, high cholesterol) cardiovascular diseases. Lastly, it will collaborate with the industry to co-develop novel AI models for 3H patients.

4. Personalised education through adaptive learning & assessment

The government will focus on transforming the education system, particular for the primary and secondary levels. Here, it aims to develop adaptive learning systems and help students learn via AI, and also to implement automated marking systems to aid teachers.

By 2022, the government plans to launch automatic marking systems for the English language, to primary and secondary schools. By 2025, this will be extended the mathematics for primary and lower secondary levels, and, by the same year, a Learning Companion will be launched to give students personalized coaching. Finally, by 2030, these marking and learning systems will be expanded to include more subjects.

5. Border clearance operations

Travellers rejoice! Singapore will be implementing 100% automated immigration clearance for all travelers by 2025, which means no more queuing up to get your passport stamped. Before this happens, however, the government will focus on border security, and will launch an AI-support forward risk-assessment capability by 2022.

The $200m supercomputing upgrade

The strategy’s only mention of the government budget relates to a $200m investment in supercomputing infrastructure, to be made over the next 5 years. Here, the government intends to improve capability and network speed and quality, and raise supercomputing resources from 1 petaflop to 15-20 petaflops, to support high-end compute performance needs.

In addition to these, the National AI Strategy also elucidates the government’s role in policy, such as developing sector-specific AI governance frameworks, developing training in AI ethics, and managing potential risks and problems relating to AI.