Author: Sutapa Amornvivat, Ph.D.
Published in Bangkok Post newspaper / In Ponderland column 20 September 2017
A few weeks ago, a coalition of labor unions demanded doubling Thailand's minimum wage from currently 300 baht per day to 700 by 2018. The proposed number aims to meet a living cost, high enough to support two family members, assuming an inflation rate of 3% a year.
Whether this policy is good for the Thai economy is up for debate. The issue of minimum wage has always been social as much as, if not more than, economic. And its unintended consequence to the marginalized groups has already begun.
We can look to a recent economic research by Grace Lordan and David Neumark for a glimpse of what may come to our economy. The research finds that, increasing the minimum wage could drive low-skilled workers out of jobs faster, replaced by machines. The impact is particularly strong for the most vulnerable, in this context, the elderly and minority in the U.S. This is an example of how some public policies can unintentionally tilt the balance in favor of machines, as humans and machines compete for the same jobs.
With emerging Artificial Intelligence (AI) technology, machine can now self-learn from data and complete tasks like human. It has started to occupy activities that people used to put time and effort in, to now go on autopilot mode. The anxiety against machines that has been around since the industrial revolution has re-emerged and intensified due to the rapid development of AI.
Even without a push from this policy, Oxford University researchers already predicted that almost half of U.S. jobs are at risk of becoming automated within the next two decades. Moreover, a report by the World Economic Forum, "The Future of Jobs," also sees a trend with AI taking over the world at the start of the Fourth Industrial Revolution. The implication is an overhaul to labor markets globally, as firms begin to weigh costs between using machines and human workers to do the same jobs.
In fact, we are already seeing price of machines go down while cost of hiring workers are generally rising in the United States. Bank of America Merrill Lynch reported that the cost of using a welding robot is three times cheaper than hiring a worker on an hourly basis. At the same time, the cost of building one has dropped almost 30 percent over the past ten years. The cost reduction will not slow anytime soon, as these machines continue to become more efficient.
On top of getting cheaper, machines are now getting better at a wide variety of tasks. The scope of "automatable" jobs is fast expanding. In the past, automatable jobs usually involve physical routine tasks such as manufacturing assembly lines. But, breakthroughs in AI and machine learning now allow for automation of more cognitive tasks such as financial analysis, insurance underwriting, and sales. In the future, automation can expand to truck and taxi drivers, doctors and even a creative job. Indeed, Amazon is building a fashion designer AI by having machines learn from images posted on Instagram and Pinterest.
Although some of these machine-competing jobs may seem far-fetched for Thailand, our path to automation may take place sooner than expected. According to Thailand's Board of Investment, Thailand's imports of automation industry rose 8% a year from 2010 to 2014. At present, most of these machines remain in automotive, electronics, food processing factories, where tasks are more physically routine.
The pace of shifting towards automation could pick up as we too are facing rising costs of human workers. The minimum wage hike to 300 baht per day in 2013 caused a 60 percent rise in labor costs in real terms when it was first introduced.
In addition, our population is rapidly aging, and workforce shrinking. Immigrant workers, the go-to alternative for Thai firms, are also facing a tight registration control. These phenomena will soon push the cost of human workers up to the point where automation becomes a viable option to explore.
In fact, an interview of auto makers in Thailand, by the International Labour Organization, refers to the wage hike in 2013 as a "wake-up call to integrate automation." Electronics makers in Malaysia share similar story as the country introduced a minimum wage in the same year.
With this in mind, raising minimum wage is unlikely to solve the problem of economic disparity. On the other hand, the policy runs a high risk of leaving some workers further behind. Minimum wage increase without improving productivity can outright hurt the very people it tries to help.
A more sustainable remedy is to focus on training or, more appropriately, re-training people to leverage machines in delivering a task.
Today, free training courses are available online. A sales clerk can learn to become an online marketer; a factory worker can learn to write computer codes. Whatever prerequisite skills required are accessible on the internet as well. Guidance in navigating these online contents and perhaps making them available in Thai language could be a combined effort between firms and the government to ease the transition.
At the frontier of technology, we also see a movement among researchers and innovation leaders advocating for development of AI that complements rather than replacing people. Barbara Grosz, a professor of Computer Science at Harvard University suggests that "human and machine capabilities can add up to more than the sum of their parts." This is echoed by Microsoft CEO Satya Nadella who say that his company aims to build AI to assist human, not replacing as well. A research group at Google, "People + AI Research (PAIR)" is dedicated to making technology transparent and design AI that incorporates human interactions.
It is likely machine will replace humans in many jobs, disrupting various industries. But it is far from becoming a dystopian future that some has paint out earlier. All things considered, AI can free people away from undesirable tasks – ones that are dirty, dangerous, and repetitive. It will give us the chance to find things we truly care about.
So, rather than holding on to fears of being replaced by machines, we should concentrate our energy on finding a way to leverage them, bearing in mind of side effects of well-intended policies. A safety net in the form of education programs will do a better job at ensuring a stable economy and an inclusive society in the long run, where machines can work alongside us in achieving more together.