Protecting workers before AI widens digital divide

WHILE artificial intelligence (AI) is all the rage these days, there is no need for Singapore to “rush headlong” into AI without first understanding its impact on workers.

Such was the message from Prime Minister Lawrence Wong at an SG60 conference on Jul 29. Amid fears of AI replacing jobs, especially entry-level ones, the prime minister’s nuanced take is refreshing and should offer some reassurance. He said pioneers of cutting-edge technologies do not necessarily become industry leaders, and that economies only reap the real benefits of a technology when there is broad-based adoption – the latter is what Singapore should pursue.

Yet, PM Wong also acknowledged concerns about jobs. Even if every new wave of technology over the millennia ends up killing some jobs and creating new ones, history may not guide our future, he said, given how much more powerful AI is. “I would say that even as we think about broad-based adoption of AI – which we have to do because we have no choice, we have to harness technology – we also have to think equally hard about applying technologies like AI in a meaningful and deliberate manner that creates jobs for Singaporeans,” he said.

Last month, a piece in the Financial Times similarly advocated being a “second mover” in situations of high uncertainty and risk, arguing that “what you lose in speed, you gain in information”.

This presents a difficult balancing act. While a deliberate strategy is wise, the ground is already shifting as global companies rapidly deploy AI. Therefore, the need to protect and reskill workers is an urgent response to a transformation that is already underway. The force of AI is unstoppable, particularly among multinational companies, many of which also operate in Singapore. For example, over three quarters of McKinsey employees now use its proprietary AI tool to create PowerPoint slides, Bloomberg reported in June.

As AI tools produce work once done by junior staff, Spanish economist and politician Luis Garicano suggests AI could destroy the economics of professional development.

He argues that companies have long relied on entry-level roles to function as a self-funded training ground, creating a sustainable talent pipeline. By automating these tasks, AI dismantles the system. This creates a “supervision threshold”, where workers skilled enough to supervise AI thrive by leveraging its power, while those below this threshold compete directly with it, facing wage pressure and obsolescence.

In other words, if AI is allowed to cut entry-level roles – which it is starting to do – we could end up with a profound digital divide.

In April, a United Nations report sounded a similar warning. While AI is on course to become a US$4.8 trillion global market by 2033, its benefits may remain in the hands of a privileged few “unless urgent action is taken”. Up to 40 per cent of global jobs are at risk, the report estimated, but government investment in reskilling, upskilling and workforce adaptation may help enhance employment opportunities.

Garicano said any viable solution would need to satisfy three economic constraints: first, compress the time required to reach the supervision threshold; second, public policy must create new incentives for private-sector investment in training; third, prepare workers for the possibility that the supervision threshold will continue rising.

To that end, it is encouraging that the labour movement is refreshing its Job Security Council in response to AI, but the devil, as always, is in the details.