Will artificial intelligence disappoint?

 

Talent Corp’s survey of the impact of artificial intelligence (AI) and the digital and the green economy clearly shows the limits and challenges of technology in the future of work and raises questions on whether so much government time and money should be devoted to it.

This “landmark” study actually pushes a storyline devised in 2013 by Frey and Osborne, which posited the alarming and wholly untrue claim that 47% of jobs in the US would be lost due to automation.

Multiple consultancy groups have pushed this fallacy since then even though Frey and Osborne themselves now advocate a more nuanced interpretation that automation, digitalisation and AI influence tasks, not jobs, and unfortunately we still need accountants even though most of their dull, routine work will be automated.

In the headlines, Talent Corp tries to create the same startling impact and sense of urgency for AI investment and training, of course all funded by the taxpayer and run by, you guessed it – middlemen, agencies and civil servants.

The patronage cascade is clear, with 10 money pots for 326 recommended initiatives, including 114 (35%) for the government, 80 (25%) for industry players, 59 (18%) for academia, and 73 (22%) for training providers.

However a look behind the numbers reveals a different story. The headline proclaims 620,000 “jobs” will be highly impacted within three to five years in 10 sectors covering 60% of GDP. This is only 18% of the jobs in these sectors and only 3.6% of the total labour force.

More than 72% of the jobs at risk are in wholesale and retail trade and food manufacturing and services – in other words online shopping will hit retail and food outlets. This is hardly a cause for alarm or a justification for urgent government interference at the expense of the taxpayer.

The good news is that 96.4% of jobs will not be highly impacted and 7.2 million people outside of the workforce will also not be affected. The bad news for these people is that the government advisers do not seem to care about them and are chasing tech-bro strategies instead.

The reality is that as more Malaysian companies announce AI related products and business models, most are basic AI adoption such as integrating Chabot products developed overseas.

Malaysia has historically been a technology adopter rather than a technology creator. This reflects the absence of R&D investment as well as the comparative advantage of leveraging technologies created elsewhere in application to local business needs.

There is nothing wrong with this as long as it adds value. Often AI replaces people which causes unemployment and underemployment. So there is a balance to be struck in the cost-benefit trade-off.

Where AI can enhance the efficiency, productivity and quality of businesses and at the same time support employment then you get a better outcome than using technology that causes social costs through raising unemployment.

The rate of adoption is in line with the demand for AI driven applications and the potential business benefits. The market should drive AI adoption and it should not be pushed for its own sake. There must be a business case.

For most Malaysian businesses, especially MSMEs, there is no obvious business case yet so pushing AI could distract companies from perfectly successful business models in search of unproductive and disruptive change.

It will also distract spending from other priorities in the care economy for example which actually need support for training and skills development.

The government should free the economic environment to allow market-driven access to AI and let businesses choose for themselves. If there is a business case then entrepreneurs will follow the benefit trail. - FMT

The views expressed are those of the writer and do not necessarily reflect those of MMKtT.