AI will affect jobs unevenly. Singapore’s response must be just as targeted: Opinion

Will artificial intelligence take away our jobs?

Elon Musk has described AI as a “supersonic tsunami” that could rapidly replace computer-based work, with AI and robots eventually making employment optional.

Singapore’s parliamentary debate in May on avoiding “jobless growth” is therefore timely. AI adoption is essential for competitiveness, but its impact on workers cannot be left entirely to the market.

A Ministry of Trade and Industry (MTI) study released in August found that early AI adoption was associated with higher firm revenue and total employment, not immediate job losses. But the gains were uneven across firms, sectors and workers.

This points to the real challenge. AI transformation will involve both job creation and job destruction, but the two effects may not move in step. Displacement can come early, while new jobs may emerge only after firms redesign work and build new capabilities.  

That doesn’t mean we slow AI adoption to preserve existing jobs. But we need to move beyond asking whether AI is good or bad for jobs, and focus on three practical questions: Which AI? Which workers? Which sectors?

Which AI? 

AI is often discussed as if it were one technology with one effect on jobs. But there are two versions of it. 

The first is cognitive AI: large language models and software agents that can draft documents, write code, conduct research, analyse data, among other capabilities. Its impact is most quickly apparent in office work.

The second is embodied AI: intelligence built into robots, vehicles, machines and physical infrastructure. Its impact will be felt more in blue-collar jobs in factories, warehouses, ports, construction sites, security and transport.

This distinction matters because the labour market risks are different.

Cognitive AI may not wipe out entire professions. Singapore will still need lawyers, accountants and software developers. But firms may need fewer people to perform the same volume of work, especially at junior and intermediate levels.

That is a danger. The tasks most exposed to AI such as research, basic coding and financial analysis are how young professionals learn. By doing the groundwork, they acquire context, recognise patterns and develop judgment.

If AI takes over these tasks too quickly, firms may still need senior professionals but hire fewer juniors. The career ladder will be weakened at the bottom. 

Embodied AI presents a different scenario. In many physical sectors, Singapore’s challenge is not excess labour, but too little of it, especially for work that is repetitive, physically demanding or risky.

Here, automation is less about replacing workers than keeping such economic activities viable. Without higher productivity in these sectors, more operations may become dependent on foreign manpower, move out of Singapore, or disappear altogether.

Which workers? 

AI will not affect all workers in the same way. 

Young workers may feel the squeeze first. Graduate employment has softened, with the share securing full-time permanent jobs within six months falling from 87.5 per cent in 2022 to 74.4 per cent in 2025, while those receiving no job offers rose from 4.1 per cent in 2023 to 8.5 per cent in 2025.

AI is not the only reason, but there is a correlation: generative AI is becoming mainstream just as entry-level hiring is weakening. 

Mid-career workers face a different pressure. Their jobs may not disappear completely, but fewer people may be needed to do the same work in the long term. Yet they could also be among the first to benefit if AI amplifies tasks involving judgment, analysis, management and client interaction.

There is already evidence supporting this. MTI’s study found that initial employment gains from AI adoption were concentrated among local higher-earners and mid-career workers, likely because their roles contain more AI-complementary tasks. 

But the good news is as firms deepened their AI capabilities, local employment gains broadened across wage and age groups. It shows that AI’s benefits will spread if adoption is matched by job redesign and worker training.

However, if those measures are not implemented, AI could still widen inequality even without causing mass unemployment. Those who can use AI well will become more productive and valuable.

Those with weaker access to training, lower mobility or fewer AI-complementary tasks may face weaker bargaining power and narrower job options.

Which sectors? 

AI’s impact will depend on not only what the technology can do, but also where it is deployed.

In digitally intensive sectors such as finance, professional services and software development, AI can reshape work quickly because many tasks are already digital. The first sign of its effect may not be retrenchment. It may be slower hiring, smaller teams and fewer replacement roles. 

The consideration is different in physical sectors, like construction, logistics and advanced manufacturing, where the issue is often persistent manpower shortages.

This is especially salient for advanced manufacturing. Factories in China are already operating with very high levels of automation. Some are moving towards “lights-out” production with minimal human intervention. 

If Singapore wants to remain a key node in the global manufacturing value chain and compete with factories in China, robotics and embodied AI will be essential to keeping production lines competitive and anchored here.

They can reduce foreign manpower dependence while opening opportunities in robot maintenance, systems integration, robot training and autonomous fleet management for local workers. 

A differentiated Singapore response

Singapore should therefore be watchful where cognitive AI risks hollowing out career pathways, especially for young and mid-career workers.

But we should be more ambitious where embodied AI can overcome manpower constraints, strengthen competitiveness and keep critical economic activities on the island.

There are four things we should do.

First, protect the first rung of the career ladder. The government, educational institutions and employers should jointly redesign junior roles in highly exposed sectors such as finance, law, professional services and software development.

Structured apprenticeships, work-study pathways and supervised project assignments should expand so young people continue to gain practical experience even when AI performs much of the traditional groundwork.

Second, accelerate embodied AI where labour is scarce. Singapore should aggressively deploy robotics and autonomous systems in sectors facing manpower shortages.

A differentiated foreign worker levy could sharpen the incentive to automate where robotics is viable. If machines can perform the tasks, higher levies for these roles can make investment in robotics more attractive than continued reliance on lower-cost foreign labour.

Trade associations and large anchor companies can aggregate common needs, procure shared solutions and spread implementation knowledge among small and medium-sized enterprises (SMEs).

Third, tailor support to workers at different life stages. Generic AI training will not be enough. Young workers need entry opportunities, mentors and workplace exposure, while mid-career professionals need conversion pathways tied to actual vacancies, employer attachments and recognition of existing skills.

Fourth, make enterprise and workforce transformation inseparable. Businesses should treat AI as organisational transformation, not just an IT project or a cost-cutting tool.

MTI’s study shows AI adoption builds on prior investments in data, cybersecurity, business software and digital workplace tools, and may take years to translate into productivity or profit gains.

For SMEs, support must therefore go beyond AI subscriptions to help them prepare this infrastructure. And productivity gains should be shared through better wages, skills allowances, career progression or improved working conditions.

A compact for the AI age

Avoiding “jobless growth” does not mean preventing every job loss. It means pursuing productivity while protecting livelihoods. That requires helping workers gain experience, build new capabilities, move into viable work and share in the technology’s gains.

Singapore’s advantage is our ability to act quickly across government, businesses, unions, educational institutions and trade associations.

If we distinguish where AI creates risks and opportunities, and act early, we can secure both competitiveness and good jobs.

The writer is CEO of the Singapore Business Federation 

Source: The Business Times © SPH Media Limited. Permission required for reproduction.

A ‘human in the loop’ isn’t enough when AI starts acting for us: Opinion Singapore’s asset management ambitions face next test as industry seeks speed and clarity