Blog

[SINGAPORE] “Jobs, jobs, jobs” – that is Singapore’s No 1 priority and economic strategy.

“I assure you, even as we embrace AI (artificial intelligence) and technology, we will not lose sight of our key priority,” Prime Minister Lawrence Wong proclaimed in his 2025 National Day Rally speech. 

Amid reskilling opportunities such as the SkillsFuture Level-Up Programme for Singaporeans aged 40 and above, and a slew of courses offered by institutes of higher learning for workers to upskill, are some job roles and industries more vulnerable than others in the unstoppable march of AI?

Which jobs AI will affect most – and least

Economists, policymakers and AI experts agree on one thing: All industries will be affected by AI. Even manual jobs, albeit to a lesser degree, said Jerrold Soh, an assistant professor of law at Singapore Management University (SMU).

“A main business line, such as plumbing, could be less affected – but auxiliary yet important functions like human resources and finance will” feel the impact, he said.

This means that even industries perceived as related – or parent – to AI, such as software and technology, will not be spared.

“While it seems ironic, this is due to (the industry’s) value shifting from repetitive raw coding to problem-solving, integration, architecture and other higher value-added expertise,” said Zafar Momin, an adjunct professor at National University of Singapore Business School.

However, the bottom-line assumption that AI will replace repetitive and predictable work still holds, he noted. Clerical, administrative jobs are likely to take the first hit.

Professions that rely on a “sizeable amount of junior staff” for data collection and analysis – as seen in auditing, legal and consulting companies – could undergo a significant number of job displacements, too, said the academic.

But young adults such as Timothy Goh, an auditor at a multinational firm in Singapore, are sanguine about AI in the workplace. “I don’t feel threatened by it – it is simply something I will need to learn because my firm is applying AI in its work and so are my clients,” the 27-year-old said.

Adam Harper, managing partner at Ashbury Communications, acknowledged that the existing business models of public relations agencies depend heavily on junior employees working long hours on repetitive tasks.

But he also noted: “If agencies can achieve efficiency gains by automating more manual processes through AI, there is potential to reduce the risk of burnout and offer junior employees more time to acquire higher-value skills, like strategic thinking or data analysis expertise.”

Liew Nam Soon, EY managing partner for Asean, pointed out that productivity gains do not directly translate to a reduced headcount. “It is more about shifting labour towards other tasks to effectively address new demands.”

The technological advancements with AI ultimately create new jobs and demand new skills, he added.

For instance, EY had 2,100 technologists in 2018, but now employs more than 80,000 technology professionals worldwide – with roles such as mathematicians, AI engineers, product managers and cloud professionals emerging and functioning alongside those in consulting, law, sustainability and auditing.

“AI skills are also no longer confined to developers or IT teams or technology professionals; instead, all EY employees pick up AI literacy skills, with strong interest and take-up rates reflected,” he said.

At EY’s tax department, AI helps to automate tax compliance processes and predictive analytics, while AI models assist in tax planning by forecasting future tax liabilities and identifying optimisation opportunities.

In March, the company launched its EY.ai Agentic Platform, with its initial phase deploying 150 AI agents to assist 80,000 EY tax professionals with data collection, document analysis and review, as well as income and indirect tax compliance.

Liew believes that there continues to be space for both human workers and AI to function alongside one another, for more productive goals to be achieved. “AI is not replacing jobs but simply reshaping them, especially in the field of professional services.”

AI isn’t foolproof

AI users and experts also note that certain models are not ready to run on their own in replacement of a human worker.

Goh, the auditor, said the AI tools he uses now can be complex to navigate, and can hinder his traditional workflow. “I understand that it’s meant to make our lives easier – but a learning curve still exists. AI isn’t infallible, (either), which means we still have to review the output from the tool.”

This issue extends to other sectors such as legal work, which has very little margin for error. “A single missed typo in a contract could end up in a million-dollar dispute,” said SMU’s Prof Soh.

He added that lawyers who use AI tools will need to worry about missed “hallucinations” – false or misleading AI-generated information presented as facts – especially in client or court documents.

“The issue is tricky… lawyers will need to become adept at spotting things like fake case law and misquoted documents in a way that allows them to meaningfully use AI without spending a disproportionate amount of time correcting its errors,” he said.

AI has also produced a generation of Google-informed clients who are frowned upon by medical and legal professionals. Prof Soh cited examples of clients presenting their lawyers with ChatGPT-generated legal drafts, asking them to simply build on this to “save costs”.

“This may end up making things more costly as the lawyer has to first check and correct the AI’s mistakes before customising the draft for a client’s unique situation,” he explained.

Efficacy and cost-savings aside, Ashbury’s Harper believes the human touch is still very much necessary for work involving creativity and strategic thinking.

He pointed to how LinkedIn users quickly identify and reject shallow and “clearly AI-generated” content that appears on the social media platform.

“This really highlights the need for humans working in content and communications to direct, edit and refine the creative process to ensure that the content connects with their audience,” he said.

Will history repeat itself?

The main concern now for workers and jobseekers is the possible mismatch between the number of new AI-driven jobs and those being replaced by it.

The issue, however, is not a new one, said Jacob Mok, an associate faculty at the Singapore University of Social Sciences. He noted that this was also widely discussed when technology revolutionised the job market.

“In the past, industrial automation led to early fears (of worker displacement)…. It didn’t happen – at least to the extent feared – because new job roles emerged in the process,” he said.

“While there is a case for AI-based automation and output posing a more formidable challenge to today’s workforce, I wouldn’t be surprised if this concern – just like with similar, job loss-related concerns from yesteryear – turns out to be overblown.”

Prof Soh also cited how AutoCAD, a design software, has replaced pencil and paper blueprints – yet this has not affected the way architects thrive today. Similarly, spreadsheets evolved from ledgers for accountants.

“Those who pick up new tools can do new things, too, within the same industry,” he said.