No career is safe. Every industry, from manufacturing to medicine, is at risk of a robot taking over.

“It’s been proven statistically that a machine can analyse a tumour and read it just as well as a radiologist can,” says Andy Zook, ASEAN Vice President of global business analytics provider SAS. He adds that reskilling is much more important now, with the rise of artificial intelligence (AI) making a “big impact on industries that for decades were great jobs, and now are at high risk”.

Governments now have to think about “how to shape workforces”, Zook says, especially when healthcare providers may need to only hire one-fifth the number of radiologists in the future. “The US is already seeing that radiologist employment is flattening, and it’s predicted to decline,” Zook notes.

“Old models are not working; new models are coming thick and fast; and we are having to adjust and to keep up because of technology and globalisation,” said Singapore Prime Minister Lee Hsien Loong during the country’s National Day Rally last year. The government must help older businesses adapt to the new business models, he added.

Predicting what lies ahead

Analytics is one way that governments are “forecasting sunset and sunrise industries”, says Lee Tian Lee, General Manager, Singapore, for SAS. “If there are certain industries that are going to be ‘sunsetted’, we better make sure people don’t go into it,” he says. SAS is currently working with the two government agencies promoting lifelong learning, SkillsFuture Singapore and Workforce Singapore (formerly the Singapore Workforce Development Agency), to carry out this forecasting for industries in Singapore.

For workers in a sunset industry, their skills may be transferable to other sectors, and analytics can help “to matchmake workers and employers together”, says Lee. This will involve text analytics and taxonomies of relevant words. “From the job posting and job descriptions and resume, you can start to do this matching,” he explains.

With these insights, employers and governments can start training and preparing workers for sunrise industries earlier, and foresee the need for certain skills four to five years ahead, Lee adds.

Sniffing out fraud networks

Identifying tax fraud is another area where analytics prove useful, according to Zook. Governments will be able to predict the “likelihood that a citizen might avoid taxes”, and identify “suspicious behaviour” that help them uncover fraud networks, he explains. “As the sophistication of criminals rise, the techniques and the tools that governments and financial services firms use have to rise with it,” he notes.

To do this, the technology will pick up on sets of “network relationships” or “networks of behaviour”, Zook says. “If you think of social networks, different people know each other through various degrees of separation. Those same concepts come into understanding how people are avoiding taxes,” he explains.

This trend of using analytics to detect fraud is catching on. In Singapore, a tool built by the Agency for Science, Technology and Research (A*STAR) alerts officials to suppliers and employees involved in suspicious deals. The country’s Ministry of Finance also uses machine learning, a type of AI, to pick out financial anomalies and flag them for audit.

Data science in government

Governments themselves are facing disruption as well, which is affecting the kind of skills they need. “They need civil service leaders to appreciate data science and know how to consume analytics to make better decisions,” says Lee.

The data scientists of the future “need to have some competency in the AI space”, Lee says, especially as Singapore works towards its vision of a Smart Nation. Singapore is building a nationwide sensor network, which will collect vast troves of data that require “real-time analytics” to make sense of, he says.

AI can augment the work that these data scientists do, as it “is really good at solving a problem that has a finite number of steps and can be repeated over and over again,” according to Zook. “Data scientists can say, ‘OK, AI algorithm, you go find that stuff while I work on solving the rest of the problem’,” he says.

In practice, AI can be used to dynamically change road pricing systems to optimise traffic flow, Lee explains, as this task does not need a human anymore.

Governments now have various tools and technologies available to them to help them detect fraud in taxes, finance and procurement.

Technology comes with the added benefit of freeing up the humans to do high-level work, as algorithms chug away in the background, analysing patterns and anomalies.

And while AI will mean the fall of some jobs and perhaps entire sectors, it’s important to remember that it can also help identify what skills will be needed in the future.