W
hen the Commonwealth Bank of Australia replaced 45 customer service representatives with a voice bot in July, it looked like the future had arrived.
"Our investment in technology, including AI (artificial intelligence), is making it easier and faster for customers to get help, especially in our call centres," the bank said in a statement to a news organisation in Australia.
But what happened after that was far from what the bank and many were expecting.
Instead of reducing wait times and improving service, more calls poured in. Customers grew frustrated, complaints mounted, and within weeks, the bank was scrambling to rehire the very staff it had let go.
The misfire offers a cautionary tale: even in sectors where automation seems obvious, AI can falter when empathy and human nuance matter. It also underscores a middle ground Malaysia should be preparing for.
“(Malaysia’s) approach must be human-centred automation – AI should augment, not replace people. We need to design workflows where AI handles scale and repetition, while humans do what machines cannot: empathise, interpret complex context, innovate and uphold values,” said MBSB Bank chief people officer Farid Basir, who is also a council member of the Malaysian Employers Federation.
CLERICAL JOBS ON THE LINE
In Malaysia, jobs in the clerical and administrative sector are the most vulnerable to disruption. A joint study by the Institute of Strategic and International Studies (ISIS) Malaysia and the World Bank estimates that 92 percent of clerical support roles are highly exposed to generative AI.
World Bank economist and co-author of the study, Alyssa Farha Jasmin, told Bernama their research does not predict inevitable layoffs but shows how vulnerable these roles are to automation. Exposure means that their tasks overlap with AI and could be automated, depending on how employers deploy AI.
The overlap could affect some 4.2 million workers in Malaysia – nearly a third of the national labour force – whose daily tasks, from processing records to preparing invoices, can now be performed at low cost by machines.
Women are likely the most affected as 84 percent of clerical staff, such as receptionists and secretaries, are female. The study also lists younger workers (aged 25 to 34) as being vulnerable to AI, despite having lower exposure than older workers.
“We find that the youngest workers (aged 15 to 24) are actually most represented in the lowest exposure to AI (segment), but it is the prime-age workers (25 to 64) that are most represented in the highest exposure to AI. So what this means is that younger workers may be trapped in low-skill roles as they tend to have more manual roles that require little to no formal qualifications or experience, while entry-level cognitive jobs that are held by these prime-age workers can be increasingly automated,” Alyssa said.
BREATHING ROOM
Despite the doom and gloom of AI replacing workers, Malaysia has some time. While large international companies have adopted AI and cut jobs as a consequence – ByteDance’s TikTok laid off up to 500 workers in Malaysia in December last year as the social media platform depends more on AI for content moderation – most companies in Malaysia are not ready to adopt AI fully.
The Tech-Driven Sustainability Trends and Index 2024 survey commissioned by Alibaba Cloud found that 68 percent of responders believe companies are lagging in cloud computing and AI adoption. This hesitation is driven by barriers such as knowledge gaps, cost constraints and lack of technical capabilities.
The PwC Malaysia Corporate Directors Survey 2024 found that Malaysian CEOs are also not fully on board with AI, with only 23 percent planning to weave AI into their workforce and skills strategy. This highlights a crucial need to shift leadership mindset on AI and GenAI (generative artificial intelligence) for corporate Malaysia.
SME Association of Malaysia president Dr Chin Chee Seong said AI adoption remains limited beyond light use of generative tools. Many owners do not see clear business value and some even resist basic digital tools.
“They don’t even want to touch a notebook (computer). But they will ask their staff to do it. So they themselves don't, but their staff may use it. Right? And some of them, they don't even (want to) pay money. They want it for free,” he told Bernama.
He added that without clear incentives or support, most businesses will not adopt the technology at scale.
For now, that reluctance buys time. However, Chin agreed that as generative AI becomes cheaper and easier to use, the pressure on firms to integrate it will grow. He said he expects GenAI to become commonplace in businesses in about 10 years.
KEEPING THE HUMAN IN AI
Therein lies the crux of the issue. Without a serious commitment to maintaining human capital, many jobs will be lost to the detriment of the nation.
Lead author of the ISIS-World Bank study Calvin Cheng said the nation’s AI strategy should always emphasise technology as a way to solve human problems rather than focusing on firm or corporate profits.
“I guess in my personal view, it's about deploying technology in a way that’s beneficial for society as a whole and that doesn't necessarily, isn't always in line with firm profits as a whole, right? So we have to make a decision whether we want to prioritise firm profits or we want to prioritise societal good,” he said.
In the ISIS-World Bank paper, the authors emphasised the importance of training present workers, especially those who are highly exposed to AI, to use GenAI to augment and improve their work.
The government has pinned much of its response on upskilling and reskilling workers. Talent Corporation Malaysia Bhd (TalentCorp) has allocated RM1.6 billion to prepare Malaysians for new digital and AI-related skills, with training courses available for free for Malaysian workers.
However, companies are still not focusing enough on keeping the human element in the process of adopting GenAI.
Many surveys done on companies’ AI attitudes have found that the human-centred approach ranks low. One of the studies, EY 2023 Work Reimagined Survey by UK-based Ernst and Young Global Limited, found that most companies in Malaysia tend to underestimate the need to upskill and reskill their workers for AI, with only 22 percent preparing to provide training to their workers.
The numbers are even lower in the PwC Malaysia Corporate Directors Survey 2024. This survey found that 12 percent of corporate directors in Malaysia thought training employees to use GenAI was important, while only four percent were concerned about job losses as a result of GenAI.
Another point the study stressed was the importance of human edge skills, such as creativity, critical thinking, complex judgment and dexterity, which GenAI is still unable to replicate. (The easiest way to explain this is AI is not actually an intelligence as it is not sentient – currently, it is basically still very much a predictive model based on all publicly available data – so it cannot come up with original thinking.) Hence, skills such as creativity and critical thinking would largely insulate capable workers from being replaced.
Judges, surgeons, dentists, aviation instructors and massage therapists are among those rated least susceptible to automation as their jobs require discretion, dexterity or deep human interaction.
Farid of MBSB Bank agreed, noting that removing the human element leads to a "loss of empathy and relationship-building", which are vital in client-facing roles. He also warned of the systemic risks of over-reliance on AI, including flawed models and cybersecurity breaches.
Still, experts caution that no sector is entirely insulated, given how rapidly the technology is improving.
ONLY GOOD ON PAPER?
But not everyone is convinced. While Hema Kangeson, a consultant in organisational culture and leadership, concurred that retraining and upskilling workers are important, it would not save all workers from job loss, especially those who are incapable of adapting to technological changes.
“There is no clear analysis and data to say (retraining) is going to be successful. Everything is a finger in the air at the moment, you know,” she said.
“Companies might upscale some people, but they're going to get rid of others because if they're going to be continuously using AI, there are going to be a lot of areas where they might not need people for certain tasks anymore. There might be new jobs created as well due to AI, but it's not entry-level jobs.”
For younger Malaysians, the prospect of a shrinking entry-level job market is daunting.
Nik Nur Rasyiqah Nik Mohd Redzuan, a 22-year-old engineering student at Universiti Teknologi Melaka, said she fears that by the time she graduates, there will be fewer opportunities for fresh graduates, despite her comfort with using GenAI.
Her concern reflects a broader unease: even if graduates gain AI skills, existing biases in hiring may leave women sidelined in a tighter labour market.
“I am in engineering and have many male friends in engineering, but I notice that when it comes to internships, they mostly look for men,” Nik Nur Rasyiqah said.
Alyssa, who works at the World Bank office in Kuala Lumpur, acknowledged the student’s fears, saying that while Malaysian employers should train workers in AI skills, they must also remove barriers that keep women from staying in or reentering the workforce, such as childcare access and re-entry incentives.
“I think it's a bigger issue than simply just, you know, giving them tax breaks and things like that … we need to look at the system of care and look at where we can even out perhaps the workload between men and women and make childcare accessible, affordable to the parents,” she said.
However, SME Association’s Chin disagreed, saying that GenAI would likely increase women’s chances of getting hired as it would allow them to work flexible hours.
For all the optimism about augmentation, job losses are likely. That makes social protection urgent.
The ISIS–World Bank research recommends strengthening unemployment insurance and other safety nets so that workers have support should they be displaced.
It also suggests beginning training earlier, such as mainstreaming AI education and ethics, as well as the limits and adverse effects of AI usage, and developing human edge skills. The paper also recommends offering incentives to employers to invest in retraining workers, among others.
Hema agreed. “We need to teach children AI literacy, critical thinking and creativity,” she said.
A NARROWING WINDOW
For now, Malaysia still has breathing space. Businesses remain slow in adopting AI, government initiatives are still being structured, and wages remain low enough that replacing humans with machines is often not economical.
But the window is narrowing. The ISIS–World Bank study notes that AI adoption is spreading five times faster than mobile phones, with capabilities advancing at a pace that outstrips previous technological revolutions.
The challenge for Malaysia is to ensure that this wave of change lifts more people than it leaves behind — by preparing workers, addressing institutional biases and shoring up social protections.
“Malaysian workers need not fear the AI wave. Instead, let us surge ahead — learning, adapting and reclaiming the new roles this transformation will create. With leadership, education and empathy, we will shape a future where AI liberates human potential rather than diminishes it,” said Farid.
