Singapore AI Newsletter
The field moves fast, subscribe to keep up. Previous issues here. We apologise for the lack of issues in recent weeks—we've been scrambling to meet research deadlines.
Contents
- Western Singapore will become a test area for autonomous vehicles
- Operational technology cybersecurity masterplan
- Malaysia may be developing a national AI strategy
- Sharing initiatives in Singapore
|
|
Western Singapore will become a test area for autonomous vehicles
Over the next few years, the test area for autonomous vehicles will expand to include all of western Singapore. The expansion opens up more than 1,000km of public roads. According to the Land Transport Authority, test vehicles must have a safety driver, be clearly marked, and be covered by third-party liability insurance. Road trials have been happening on 200km of public roads since 2015, at areas including Buona Vista, Jurong Island, Nanyang Technological University, and Sentosa. Trials for driverless buses are expected to begin by 2022. The timeline for level 3 or 4 autonomous vehicles remains uncertain, but whenever they arrive, KPMG thinks that Singapore is one of the countries most ready for adoption.
|
|
Operational technology (OT) cybersecurity masterplan
The Cyber Security Agency of Singapore (CSA) has released its strategy to improve the security of operational technology (OT) systems in the country. OT refers to systems that control physical processes, including traffic light controls, train signalling, and the electricity grid. A useful shorthand is that IT manages the flow of information in the office, while OT acts on machines on the factory floor. The strategy consists of four pillars: (1) training 70–100 people per year; (2) a centre for sharing threats and resources; (3) updating the existing Cybersecurity Code of Practice (CCoP) with OT-specific measures; and (4) developing new OT cybersecurity solutions. CSA has yet to release low-level details of the strategy.
Why this matters: As we shove the internet into objects (e.g. doorbells, cars, garbage cans, voting systems), cyber attacks will have much greater reach and consequences. A notable OT cyberattack is the 2017 Triton malware attack on oil refineries in Saudi Arabia. AI systems are susceptible to malicious attacks, ranging from simple (e.g. kicking down stop signs to fool driverless cars) to sophisticated (e.g. small-perturbation adversarial examples). Furthermore, using strong attacks to find worst-case inputs is one way of ensuring that models are robust to natural failures. In other words, securing AI systems against malicious attacks may also reduce accident risks.
Related: 26% of organisations in Singapore have been breached by cyberattacks in the past 12 months (Sophos). Following a few high-profile cyber attacks, the Public Sector Data Review Committee has issued its preliminary findings and is expected to release a full report in November 2019 (SG AI #3).
|
|
Malaysia may be developing a national AI strategy
The Communications and Multimedia Minister of Malaysia, Gobind Singh Deo, said that Malaysia is looking into formulating a national policy for data and AI. Previous government statements have made little mention of the social implications of AI. In this case, the minister said that "AI must be developed to include the assimilation of ethical values, taking into consideration the interests and well-being of society”. Mr Gobind said that the national policy can build on the framework that the Malaysia Digital Economy Corporation (MDEC) is due to present to the government at the end of 2019. He did not give a concrete timeline.
|
|
Sharing initiatives in Singapore (Not AI or data-related, but worth *sharing*)
The Institute of Policy Studies (IPS) has released a report looking at how sharing initiatives can build stronger and more cohesive communities. The report is broken up into two parts. The first reviews the literature on various aspects of sharing initiatives, such as their characteristics and positive and negative outcomes for individuals, societies, and cities. The second part offers insights from focus groups, identifying Singaporean attitudes toward sharing initiatives, reasons for participation, factors that encourage or hinder participation, and challenges faced by innovators of sharing initiatives.
The authors found that the majority of initiatives in Singapore help people meet immediate personal needs (e.g. convenience, cost-savings, and enjoyment) and take place at the collective level. On the other hand, community-building initiatives (e.g., encouraging civic mindedness, protecting the environment, and fostering community spirit) are less common in Singapore. The authors found that existing barriers to participation include awareness of the initiatives, cultural barriers such as the concept of ownership, and different priorities of public sector agencies that lead to less importance placed on sharing.
Jo: Two things stood out to me. First, when I hear the words “sharing initiatives”, big tech platforms like Gojek and Airbnb come to mind. The report provides a heartening reminder of other types of initiatives like community repair meet-ups and bag-share platforms that encourage the use of common reusable bags left near supermarkets (p.95). Second, it was helpful to learn more about the challenges faced by sharing initiatives in Singapore. For example, people who want to start a food sharing initiative have to navigate different guidelines from agencies like the URA, National Environmental Agency, and the Singapore Tourism Board (p.178). All in all, this report is one of the most in-depth pieces I’ve seen exploring the perspectives of those participating in sharing initiatives in Singapore and is worth your time if you’re interested and have a few hours to spare (warning, it’s 254 pages).
|
|
Others
Radin Mas Community Club touted as Singapore's first “smart” CC (CNA)
Green screen karaoke; virtual tours for room booking; facial recognition for unlocking doors. Some days I feel like Singapore is one big meme—this video confirms my suspicions.
Say Hello to the New Work Order (CSF)
This blog post about the future of work offers three scenarios centred around two uncertainties: (1) whether there will be more or less productive work for people, and (2) whether jobs will be replaced by task-based work. According to the authors, policymakers should recognise that work is tightly bound with a person's mental health, identify, and relationships. The post was originally published by the Centre for Strategic Futures in their biennial publication released earlier this year.
Algorithmic decision-making in AVs: Understanding Ethical and Technical Concerns for Smart Cities (Lim and Taeihagh 2019)
This paper from researchers at the National University of Singapore summaries the technical and policy/ethical issues related to level 4 and 5 autonomous vehicles in smart cities. The technical challenges facing are: decision making, control, and testing and verification. The policy challenges are: bias, choosing between different ethical theories, and the incentive structure of the market.
|
|
Don't keep this to yourself. Share the good stuff.
This newsletter was written by Jia Yuan (jyloke@smu.edu.sg) and Josephine (shseah@smu.edu.sg). We’re research associates at SMU’s Centre for AI and Data Governance. Malaysian updates are written by Julian Theseira. Julian is an independent researcher in Malaysia.
This week's photo: Tourists and a sleeping man at Chijmes.
Opinions expressed reflect the beliefs of individual authors and not CAIDG as an institution.
|
|
|
|
