Singapore AI Newsletter
Fortnightly updates about AI policy and the wider ecosystem in Singapore. The field moves fast, subscribe to keep up. Previous issues here.
This issue is dominated by announcements that came out of the Smart Nation Innovations Week and the Singapore Defence Technology Summit.
|
|
PDPC releases guidelines for data sharing
On the heels of the Model AI Governance Framework, PDPC has released the Trusted Data Sharing Framework. It provides non-binding guidelines on data sharing by private and non-governmental organisations. Accompanying the Framework is the Guide to Data Valuation, which lays out three approaches (market, cost, and income) that companies can take to value their data.
Jo: I was struck by two things. First, the documents normalise the imperative to collect and extract value from any form of data, bypassing discussion on whether categories of data ought not to be collected or deserve special protection. Second, the Guide to Data Valuation focuses on data sharing to increase profits and reduce cost, and pays less attention to data sharing for public good. I was hoping that considerations of ethics and social good would be more pronounced in the document.
Jia: It’s worth noting that these documents do not apply to public sector organisations, whose data sharing practices are controlled by the Public Sector (Governance) Bill. In contrast, many sections of the GDPR apply to governments in Europe; in the U.S., much of the conversation regarding face recognition has revolved around use by local governments. To whom regulation applies matters a lot, and the public sector seems to be relatively less scrutinised in Singapore.
|
|
Government-led programs to build a smarter city with 5G coverage
IMDA and the National Research Foundation have set aside S$40m to build Singapore’s 5G ecosystem, especially in the areas of maritime operations, transport, and housing. IMDA also announced a S$14m Smart Estates program, which supports projects that apply tech to urban planning and building management.
Jia: These announcements are consistent with the government-led approach of Smart Nation: the top sets the goals and strategy, and supports bottom-up solutions and proposals. One supported project aims to build no-checkout convenience stores enabled by face recognition. The company building these stores is called Le Tach—I had previously never heard of this company, and based on their website they mainly build vending machines. A reminder for researchers that their discoveries may be adopted by groups with varying expertise and intentions.
|
|
Collaboration between U.S. DOD and DSTA
As part of the Singapore Defence Technology Summit, the U.S. DOD’s Joint Artificial Intelligence Center (JAIC) and Singapore’s Defence Science and Technology Agency (DSTA) announced that they were working together to build and use AI, focusing on humanitarian aid and disaster relief. The press release says that the two organisations also “discussed possible future partnerships” that may go beyond HA/DR.
Jo & Jia: While LAWS have received the most attention, the majority of military adoption of AI is in other areas. Applying AI to HA/DR is an exciting example of digital humanitarianism, especially as satellite images and other weather sensing data become more available. Given that the U.S. DOD is a crucial supplier of defense technology for Singapore, we're interested to see what “future partnerships” look like.
|
|
Other links
Measuring Political Media Slant using Textual Data: Evidence from Singapore (Lucas Shen)
Lucas, an economics PhD candidate at NTU, uses approaches from text processing and NLP to match quotes in The Straits Times with the originating speeches in parliament. The paper finds that “the mainstream media of Singapore quotes the parliamentary speeches of the opposition politicians with less accuracy than their ruling-party counterparts” (p.38).
Second Government Bug Bounty Programme (GovTech)
GovTech and Cyber Security Agency (CSA) of Singapore are conducting a second bug bounty program to test nine government systems. The first event led to the discovery of 26 vulnerabilities and S$16K in payouts. In the past decade, such programs have become more popular in the private sector, and governments (U.S., Sweden) are starting to catch on. I’d love to see the Singaporean government run a responsible disclosure program for ML models.
Digital Industry Singapore office
The government has set up DISG to serve as the point of contact for technology companies.
Singapore Statement on Co-creating a Connected Digital Future in the Asia-Pacific
32 countries in the Asia-Pacific Telecommunity (APT) adopted the Statement, which lists five goals to guide the development of ICTs in the region. On a related note, Singapore led the ASEAN Smart Cities Network (ASCN) last year.
|
|
|
Thanks for reading! Sign up here.
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.
Opinions expressed reflect the beliefs of individual authors and not CAIDG as an institution.
|
|
|
|
