AI's March Could Take a Pause, and That May Not Be a Bad Thing

It's getting harder to gather quality data that drives AI, but the slowdown gives us a chance to make adjustments.

- Ravi Velloor

For the past 18 months, many chief executives of major companies I have spoken with, particularly in the technology field, have been sharing their assessment that 2025 is going to be something of a watershed year, when the use of artificial intelligence (AI) will become so pervasive that it will strike societies with gale force.

As automation and robotization sweep through factory floors, many of the white-collar office jobs that were thought to be safe from these trends could be under threat, or, at the very least, alter significantly as machines replace human beings, and humans and machines collaborate even more closely.

That will no doubt raise productivity by leaps; anyone who has had to spend more than an hour at a bank branch waiting for attention, as I recently did, will appreciate such work being taken over by efficient, automated intelligence.

The promise of this leap in technology, whose impact is said to rival and even exceed the arrival of the internet, is thought to be limitless.

Accenture published a report a few months ago that said fully a third of working hours across the Asia-Pacific is poised either to be automated or augmented by generative AI, leading to a productivity boost.

Working hours in Australia and Japan will be most impacted at 45 per cent and 44 per cent respectively, followed by China (33 per cent) and India (31 per cent), according to its projections.

The most impacted industries include capital markets where gen AI will transform nearly three-quarters of working hours (71 per cent) and software and platforms where two-thirds (66 per cent) of working hours will be automated or augmented. This is followed by banking (64 per cent), insurance (62 per cent) and retail (49 per cent).