As AI alters job scopes, workers and educators must make adjustments too

Through Holmusk’s GlycoLeap app, Lau’s patients send in photos of their meals. Lau then examines and rates them for nutritional value.

Instead of giving face-to-face advice, she assesses the nutrition level of the meal and texts patients with advice on how to make it healthier.

But perhaps the greatest technological disruption to Lau’s profession has yet to come — in the next few years, part of her job could be replaced by an intelligent chatbot.

Holmusk is currently experimenting with artificial intelligence (AI) to develop a bot that can make use of image recognition technology to respond intelligently to patients.

For instance, the bot could recognise a plate of chicken rice as being high in fat and advise a patient to switch to a leaner option. Lau’s job would then be to monitor the bots, while providing emotional support to patients.

According to Holmusk chief medical officer Yau Teng Yan, AI would not make Lau’s job obsolete but would fundamentally change the nature of it.

“When an aeroplane is on autopilot, you still need a pilot to monitor it... If our dieticians’ work is automated, we believe it will free them up to help a lot more patients. They can focus on the ones that need the most attention, and spend more of their time on the human touch — building rapport, counselling and providing emotional support,” he says.

Holmusk’s experiment provides a glimpse into the next frontier of job shifts: AI could automate key elements of white-collar jobs that require careful analysis and pattern recognition — from dieticians to financial advisers, and from administrators to lawyers.

“While the industrial revolution led to routine and repeatable physical labour being automated, AI is now automating [white-collar] jobs that involve mental labour as well — those that can be ‘learnt’ through careful observation and mastered through practice,” says Manjunath Bhat, a research director at Gartner.

So what does this mean for professionals like Lau, and what can they do about it? While AI will not wipe out jobs in “one fell swoop”, Bhat reckons it will challenge workers to adapt to “new ways of work” and a more dynamic labour market.

AI to benefit service-driven economy
AI encompasses a broad set of technologies that enable computers to sense and comprehend data, and thereafter take action, while adapting constantly to new information.

These technologies include the language processing and image recognition tools being tested by Holmusk, as well as analytics software, robotics and driverless cars.

With Singapore’s investment in deep technology and its Smart Nation drive, the country’s economy is poised to benefit from AI disruption. According to a July report by Accenture and Frontier Economics, AI could increase labour productivity in Singapore by 41% and nearly double the country’s annual economic growth rate from 3.2% to 5.4% by 2035, contributing US$215 billion ($290 billion) in gross value added.

A booming economy does not always mean more jobs, though. Even now, as Singapore’s GDP is beating expectations, the employment outlook remains dim.

The preliminary estimated seasonally adjusted unemployment rate for Singaporeans was 3.3% in June, down from 3.5% in March. But total employment fell for the second straight quarter.

Observers blame structural shifts in the labour market, which have led to a mismatch in skills. “Singapore is encountering a talent crunch in many sectors,” says Tatiana Ohm, vice-president and general manager of talent management firm Kelly-OCG in Southeast Asia.

Will demand for talent grow or lessen as AI becomes more prevalent in the workplace? New York University economics professor and Nobel laureate Michael Spence says AI poses less of a threat to Singapore because the technology displaces more manufacturing than service jobs.

“I don’t think we should really be afraid of technologies in a dynamic economy like this one that’s based on services,” Spence said at a press conference here last month in conjunction with the Singapore leg of UBS’ Nobel Perspectives Live! Event. “In terms of employment especially, [AI] is not the threat it might seem to be.”

Perhaps the more important question is whether the job mismatch will intensify.

As a “very high-income, advanced” country, Spence says, Singapore’s key task lies in ensuring the right match of jobs and skills. “The challenge is to engage in a process in which, over time, you figure out what the composition of jobs and employment is going to look like,” he says. “You [have to] experiment and figure out ways to help young people prepare themselves in terms of creativity, flexibility [and] continuous learning.”

‘Upskilling’ professionals
The impact of AI on skilled labour will vary across professions, says a spokesperson for the International Labour Organisation. The ILO lists accountants, sales executives and office administrators as facing the greatest risk of displacement. This is because AI is not only much faster at analysing large amounts of information, but can also retain more data with smaller margins of error.

Oversea-Chinese Banking Corp, for one, is already developing AI solutions that can help compliance analysts detect money laundering and terrorism financing.

On July 19, the bank announced that it was working with two fintech companies — BlackSwan Technologies and Silent Eight — to roll out software that can flag unusual transactions in a matter of minutes. Traditionally, it would take a compliance analyst between an hour and a few days to carry out such research.

Others in sectors such as healthcare and education are likely to have only parts of their job taken over, which will help them focus more on “interpersonal tasks” instead, the ILO spokesperson says. Echoing this sentiment, Gartner’s Bhat says professionals will have to be willing to let AI take over parts of their jobs so that they can move up the value chain.

This would mean that even professions such as lawyers and doctors could fundamentally change.

Corporate lawyer Stefanie Yuen-Thio, who is joint managing director of TSMP Law Corp, cautions that AI could replace the entry-level legal functions of junior litigators. For instance, JPMorgan Chase & Co has created an AI programme called COIN for commercial loan contracts, while ROSS Intelligence, built using IBM software, can conduct legal analysis.

Yuen-Thio believes that law firms should embrace AI innovations. Lawyers, meanwhile, need to move upstream to provide more specialised services that are not replicable by AI.

“Lawyers who want to survive will need to have highly honed human skill sets — advocacy and negotiation — or be at the forefront of developing technology and law,” she says.

In the medical profession, Gartner’s Bhat says AI could be more efficient at analysing the results of test and scans — disrupting the work of specialists such as radiologists.

“Computer vision technologies are becoming adept at analysing images, so if a radiologist is not adding value beyond interpreting medical images, it’s easy to imagine the bulk of a radiologist’s competency being driven by AI,” he says.

Holmusk’s Yau, who worked as a diagnostic radiologist for three years, has observed that more hospitals overseas are adopting AI tools to help radiologists with their work.

“I think the work of radiologists and other medical professionals may be very different in the future — in a positive way that benefits our patients,” he says. “It’s only a matter of time before this happens in Singapore.”

Taking charge of the future
As AI adoption rises, people-centric skills such as emotional intelligence, communication and persuasion will become more essential. And IT skills will become an even more valuable asset.

Citing the WannaCry ransomware attacks in May, Rajiv Biswas, Asia-Pacific chief economist at HIS Markit, predicts that cybersecurity will be “one of the fastest-growing industries, creating new opportunities for IT experts”.

Yuen-Thio believes that law schools will need to emphasise problem-solving skills rather than memorising legal decisions. Also, more lawyers will need to specialise in technology.

“The world is becoming an increasingly regulated place, and the legal issues are growing more complex in tandem with technology. These factors will require more lawyers, whether to create law at the frontiers of technology or to craft solutions around the new problems we will face,” she says.

In the case of financial services, customers are likely to still want to interact with people for services that require “some intuition or abstract understanding” of certain processes, says Sam Randall, senior manager for technology and change in financial services at Robert Walters Singapore.

“In the absence of appropriate amounts of data, AI is still unable to use judgement and gut feel to reach a decision. There will always be a human escalation point for when the algorithm fails.”

Bhat of Gartner takes this concept one step further, suggesting that human interaction may even become a value-added service for chatbots.

“Technologies such as reinforcement learning ensure that the chatbot will in due course know and remember more than the human supervisor ever will — leading to human interfaces becoming a premium service, a privilege that customers have to opt in and pay for,” he says.

While she is unlikely to be replaced by a chatbot anytime soon, Lau agrees that the value of human interaction will increase with the advent of technology. While her assessment of nutritional content can be automated, there are some things such as empathy and a listening ear that cannot be replaced, she reckons.

“Food is such an emotional issue. Even if you have a bot that is so intelligent that it can read [the] implicit nuances of people’s messages, dieticians are still needed to provide human support and encouragement,” she says.

— With additional reporting by Benjamin Cher, Jeffrey Tan, Michelle Teo and Trinity Chua