This career guide was also co-authored by Tan Zhi Xuan.
The original version of this career guide is on EA Singapore's website. I've removed large portions of it, so that it's better suited for the average EA forum audience. If you're based in Singapore, you might want to read the original version instead [which will be published soon].
Epistemic status: 60% confident that AI policy is likely to be in the top 5 highest impact career pathways in Singapore. I have spent around 30 hours conducting literature reviews, interviewing a Malaysian civil servant (the Malaysian and Singaporean civil service are pretty similar), as well as soliciting feedback from 2 local AI researchers.
Working on AI risks from the perspective of policy and governance is likely to be very impactful. The purpose of this career guide is to demystify the AI policy career pathway in Singapore, and embolden some people to pursue such a career. In this career guide, we will first argue why a career in AI policy within Singapore is potentially impactful, and then offer some recommendations on how to plan for a career in this field.
I want to thank the following people for their ideas and feedback on this career guide: Loke Jia Yuan, Tan Chiew King, and Devesh Narayanan. All mistakes and opinions in this document remain my own.
3. Is Singapore a good country to have a career in AI policy?
Most of the evidence and its conclusion that shows whether Singapore is a good country to have an impactful career in AI policy (yes, we think it is so) are already pointed out in a separate piece on technical AI alignment research in Singapore. The only difference is the following evidence against having such a career:
AI policy research institutions in Singapore focus more on risks associated with future of work, inequality, privacy, and misinformation rather than risks related to regulating the competitive development of transformative AI and shifting geopolitical powers. If you think the later risks significantly outweigh the former risks, then an AI policy career in Singapore is likely not a good fit. However, some AI policy researchers believe that research into such risks can still contribute to mitigating harm from AI.
4. Key career recommendations
4.1. How to make an impact in AI policy within Singapore
There are four broad pathways you can take to make an impact in AI policy within Singapore: civil service, tech companies with AI labs, non-governmental advocacy and think-tanks, or academia.
In the civil service, as you gain more seniority, you have the opportunity to consult and advise those who are involved in the policy formulation stage, namely those in the cabinet and the prime minister. Although you may not have direct influence over primary legislation, your position gives you a lot of accessibility to those who do. Furthermore, your direct influence over secondary legislation can be helpful in covering gaps from primary legislation. This is where you can influence AI policy and make an impact within Singapore, and possibly outside the country (if your work relates to trade and diplomacy).
In tech companies with AI labs, working in the policy arm of the company gives you the opportunity to engage with various government stakeholders on AI policy. Although you may think this might go very badly as some historical examples suggest (e.g. executives testifying that cigarettes were not addictive or regulatory capture by the oil and gas industry), some AI researchers think that AI labs should be the one taking the lead in formulating AI policies rather than governments, and that they can do it in a prosocial manner.
In academia, most of your impact will come from the quality of your research and how much your research has influenced decision makers on policy. Think-tanks are quite similar to academia in that you’ll be conducting a lot of research as well, but another big part of your work will be engaging policymakers and other stakeholders about your research. This is also the case for non-governmental advocacy more broadly, which may or may not involve research, but will certainly involve engaging the government through a combination of critique and collaboration.
NGOs and think-tanks generally have a more confined job scope, whereas academics generally enjoy more autonomy. Nonetheless, all of these options enjoy relative independence from the economic and political interests of both corporations and the Government. This allows for independent critique and development of AI policy, which may be highly impactful in providing checks and balances against corporate and governmental initiatives, in case they are not fully aligned with the welfare of society as a whole. Some examples of this in other countries include the work of the AI Now Institute and the Algorithmic Justice League.
4.3. Recommended local organisations
4.3.1. Civil service
- The Smart Nation and Digital Government Group (SNDGO) is the strategic planning arm for key national digitisation programmes. The National AI Office recedes in this organisation.
- Government Technology Agency (GovTech) is the project delivery arm for key national digitisation programmes.
- The Cyber Security Agency (CSA) oversees the nation’s cyber defenses, especially over critical information infrastructure.
- Infocomm Media Development Authority (IMDA), or more specifically Personal Data Protection Commission (PDPC), are the organisations that produced the Model Framework for AI Governance, a non-enforceable ethical guideline for AI development and use.
- Center for Strategic Futures (CSF) conducts foresight work to mitigate risks for the Singapore government. It has prior engagements with the Centre for the Study of Existential Risk (CSER) and Nick Bostrom, author of Superintelligence.
- Strategy Group develops and implements the government’s strategic priorities.
- The National Research Foundation (NRF) develops R&D priorities for the government.
- Agency for Science, Technology and Research (A*STAR) conducts research in prioritised areas.
- Civil Service College develops the capabilities of civil servants.
- The Ministry of Foreign Affairs.
- The Ministry of Trade and Industry.
4.3.2. Tech companies with known AI labs and public policy divisions
- Lazada (owned by Alibaba)
- SMU’s Centre for AI and Data Governance (CAIDG)
- NUS’s Centre for Technology, Robotics, Artificial Intelligence & the Law (TRAIL)
- NUS’s Centre for Trusted Internet and Community (CTIC)
- NUS’s Centre on AI Technology for Humankind (AITH)
- NUS’s Lee Kuan Yew School of Public Policy (LKYSPP)
4.3.4. Think-Tanks & Non-Governmental Advocacy
- NTU’s S. Rajaratnam School of International Studies (RSIS)
- LKYSPP’s Institute of Policy Studies (IPS)
- ISEAS-Yusof Ishak Institute (ISEAS)
- Singapore Institute of International Affairs (SIIA)
- Non-Profit Working Group on AI (NPWG-AI)
4.4. What are the best entry points?
4.4.1. Civil service
One of the best entry points into the Singapore civil service, especially if you’re a high-achieving student both in academics and leadership, is to apply for the Public Service Commission (PSC) Scholarships. There are many kinds of scholarships offered by PSC, so you’ll likely need to spend some time investigating which scholarship is best suited for you.
Another great entry point, if you’re close to graduation or already working, is to apply for the Public Service Leadership Programme. If you’re able to successfully apply, it might be helpful to think carefully about which pathway you want to take: the generalist leadership pathway or the sectoral leadership pathway. The generalist pathway seemed to be the most prestigious but also very competitive, as you might be considered for the Administrative Service. However, if you do have a more technical skill set and thrive in specialising in a particular field, the sectoral pathway is likely to be as impactful as the generalist one. Of the six sectors you can pick, the one most related to AI is the “Information & Communications Technology and Smart Systems” sector.
4.4.4. Civil society and advocacy
There are many possible routes of entry into civil society and non-governmental advocacy. Importantly, one’s involvement in advocacy can take place alongside an existing role as a researcher or employee in a university or tech company.
For example, the Algorithmic Justice League was initially founded by a number of academics at MIT and Emory University, while the AI Now Institute’s co-founders hold joint positions as faculty at NYU and researchers at Microsoft and Google. This illustrates one potential strategy for establishing oneself as a credible advocate on AI policy: first building career capital and a professional network within a more traditionally respected organization such as a university or a tech company, then using that as a platform for civic engagement and advocacy. Alternatively, if there are already established NGOs advocating for important AI policy issues, one can choose to join them directly.
At present, this space is fairly neglected within Singapore. While there are plenty of think-tanks that address AI policy, to our knowledge there are no other NGOs serving as more critical public advocates on AI issues. Given the success of some NGOs on other issues of public concern (e.g. AWARE’s policy achievements on behalf of gender equality), the formation of such an organization addressing AI-related policy in Singapore may be highly impactful.