It’s a two-horse race for AI crown. Where does that leave Singapore?

The obvious niche here is the regional South-east Asian market, to which Singapore could serve as a conduit or assistant to US tech giants and their clients. The oft-lamented fragmentation and diversity of these markets, and their sparse incorporation into global data bases, provide an opportunity for an aggregator in specific domains where Singapore already has expertise, like finance or maritime trade. It could also serve as a testing and assurance sandbox, checking AI agents for compliance, safety and accuracy.

To serve this role, Singapore needs to develop and provide the “human” competencies that AI lacks, and that are becoming more important as it spreads.

In addition to the personal relationships and social networks necessary to build understanding and trust, the humanities can play a role in the nurturing of AI talent and, more broadly, prepare society to better navigate an AI-disrupted world. The required domain knowledge here will be the languages, literatures, histories, philosophies, religions, cultures and arts of Singapore’s South-east Asian neighbours.

While it will be helpful for everyone to learn how to use AI as a tool in work and life, it is also necessary for software engineers to understand the humanity of others in the AI universe and be able to communicate with them.

There are good reasons for the blending of the humanities and STEM (science, technology, engineering and mathematics) skills. It is also not a recent innovation.

The University of Michigan College of Engineering has long had a curricular requirement for “intellectual breadth”. Engineering majors are required to take liberal arts courses, some of which must be at the advanced undergraduate level, where they can “learn about human modes of thought, the structure and history of the human societies that they serve as engineers, how humans behave and interact, and how humans express their aspirations in the arts, literature and music”, making them “more flexible, creative and better able to work with diverse groups”.

While there is a reflexive tendency to be dazzled at the speed and efficiency at which AI operates, its very machine-like strengths reinforce the case for the skills and habits honed through the humanities. Critical thinking is essential if one is to thrive in the AI world – to probe for possible hallucinations, disinformation and misinformation as well as to seek out more accurate and better responses through prompts.

The Socratic method I am familiar with from the privileged venues of my Cambridge undergraduate tutorials, MBA and executive education case teaching, and the corporate boardroom, can be employed more widely at earlier stages of the educational trajectory – as it already is in the A-level General Paper.

Literature and history can also provide the same practice – for example, in deconstructing characters’ natures, motivations and actions, analysing their relationships with each other and with their environments, pondering their ethics and values, and imagining hypothetical alternate plot lines (“what if?”). Reading and writing for these subjects will improve the language, and hence reasoning, facility necessary for getting the most out of AI, and disciplining it.

Critical thinking is also important in the labour market beyond, and because of, AI. The AI-producing sector itself cannot provide many jobs, given the consolidation of its shrunken corporate hierarchies in fewer, larger entities in the home country market where innovation occurs. Such jobs are also inherently unstable, with frequent and sudden layoffs and changed hiring patterns virtually “baked into” the system, given its technological dynamism and market uncertainties.

Meanwhile AI is replacing previously cognition-intensive white-collar work in other sectors, where the remaining jobs require higher-order thinking skills. Employers are starting to complain about young employees “raised on AI” being overly dependent on its derivative knowledge base and thus unable to recognise and solve novel problems which emerge. Fewer opportunities to acquire compensating tacit domain knowledge through entry-level job experience make the ability to think for yourself and know how to learn even more crucial.

AI is commoditising some creative professions like graphic design, pop music composition and pulp-fiction writing. But it also puts a premium on creativity and original thinking, which contribute to the bottom-up employment solution of individuals in the US creating their own jobs through entrepreneurial ventures, as in personal and household services, which AI can enhance.

What are the employment implications for Singaporeans in this AI era?

First, on the technology frontier, there is no visibility, only different scenarios, on the nature, pace, extent and spread of AI-induced job transformations. This extreme uncertainty makes top-down manpower master plans unworkable, even for the leading AI companies themselves. Formal technical skills training, including by universities, will be hard-pressed to keep up with the speed of technological change and obsolescence.

Second, outside of the well-established semiconductor sector, participation in the AI-producing space by hitching our economy to the bandwagon of the leading American AI companies will not create significant numbers of stable, long-term jobs for Singaporeans. The rare Singaporean or Singapore-based foreign talent that possesses the skills these companies want are most productively and securely employed at corporate headquarters, where they can imbibe and contribute tacit knowledge to the central innovation ecosystem.

Third, participation in the AI-producing space requires Singapore to possess and provide specific niche domain knowledge advantages. It does not have the resources, scale or time to “try many things at once to see what works”, as larger countries might. Singapore’s domain knowledge advantages – including tacit and intangible knowledge and the relationships providing it – are most likely to be in its geographically determined historical role in trade, transportation, finance and ancillary services in the South-east Asian region.

Fourth, the education system, and local cultural, intellectual and ideological mindsets need to prioritise the development of critical thinking skills. Critical thinking is required in the production and use of AI, and also in nurturing individual entrepreneurial values that encourage people to “create their own jobs” in non-AI sectors, which account for the vast majority of jobs in any economy.

Finally, public policies that affect employment, entrepreneurship and citizen welfare should be recalibrated to deal with the promise and perils of AI. For example, self-employment through AI-improved productivity in services is currently hampered by reliance on low-wage foreign labour. The social safety net should be strategically strengthened to support labour market mobility and entrepreneurial risk-taking.

Artificial intelligence is a revolutionary technology that is driven by the two tech superpowers of our time. It is hard to know the shape of things to come while caught in the maelstrom of change. From early indications, it involves wrenching change in the job market, with some sectors more vulnerable than others. How well Singapore deals with the disruptions will require a parallel revolution in its economic model, social contract, mindsets, and the public policies which shape them.