Digital Technologies and the Future of Work

A Post Human Job Market?

‘Humans Need Not Apply’. This was the title of a provocative video found floating around on YouTube in August 2014. It spoke of a future where new developments in digital technologies would push human labour out of the market in unprecedented ways.

The new technological leap – exemplified by artificial intelligence (AI) and machine learning (ML) – was fundamentally different from other major technological landmarks in history, the authors of the video postulated. Human beings will soon become the new horses, they argued, rendered unemployable for no fault of their own, due to the relentless juggernaut of technological development. Just as mechanical arms pushed horses into obsolescence, mechanical minds would make multiple kinds of human labour obsolete.

Since I watched this video back in 2014, reports indicating the potentially disastrous impact of AI on the job market have grown in number and size. A McKinsey report, published in 2017, predicted that by 2030, intelligent agents and robots could replace as much as 30 percent of the world’s current human labour. Of course, it all depended on various adoption scenarios, but automation could displace between 400 and 800 million jobs, requiring as many as 375 million people to switch job categories entirely and learn new skills. The National Association of Software and Service Companies (NASSCOM) similarly predicted that 5-10 percent of existing jobs could be lost to automation and another 60-70 percent of the existing workforce would have to re-skill themselves.

In other words, we are clearly looking at tectonic shifts in the job market, leading to massive job losses. The changes wrought by automation and digital technologies will not be uniformly experienced across sectors. Entire jobs might not disappear en masse. Predictive roles and repetitive tasks that are part of current jobs are at maximum risk. An article in the Economic Times, claimed that 52-69 per cent of repetitive and predictive roles in sectors including IT, financial services, manufacturing, transportation, packaging and shipping are exposed to the threat of automation.

Yet another study points out that these changes are hardly far away; nearly 50 per cent of companies have implemented AI-based ‘solutions’ in their businesses in the name of improving ‘productivity’. Given these trends, according to the above-mentioned NASSCOM report, 9 percent of the workforce will be deployed in new jobs that do not exist today, while another 37 per cent will be deployed in jobs that require radically changed skill sets.

In the Indian employment landscape, 60 percent of our occupations have around 30 per cent automatable activities. 81 percent of predictable physical activities (such as manual loading of goods), 69 percent of data processing activities (such as working on excel sheets) and 64 percent of data collection activities (such as online surveys) are most likely to be automated. Less than 5 percent of jobs are, however, likely to be completely automated. This means that overall, around 51 per cent of total employment is likely to be automated. That is still a huge number.

Old Technologies, New Technologies

From the 1970s onwards, developments in electronics spurred the growth of computer systems. The storage capacity as well as the computing speed grew by leaps and bounds. Developments in electronics allowed hundreds of computers to be connected to each other. The new digital revolution we are witnessing is, however, fundamentally different from previous technological revolutions.

Human history has experienced three major technological moments so far. We had the Industrial Revolution in the 18th century, which involved the use of steam and the subsequent mechanization of work. This created superior tools for people who worked with their hands. It transformed the manufacturing sector and pushed millions of people from agriculture into manufacturing.

Then came the Services Transformation in the 1970s. This was driven by the development of Information and Communication Technology (ICTs). For the first time, we had machines that were controlled by a software program. These machines could be used to do several kinds of tasks by simply changing the software program. These developments created jobs for people who worked with their heads and destroyed jobs for those who worked with their hands. Basically, it created substitutes for people whose jobs involved manual tasks. This devastated blue-collar workers and oversaw a surge in economic inequality.

People with very low technical skills, or ‘unskilled’ workers, were however still needed on the shop floor. The production process needed some degree of manual labour. Most jobs were lost to technology in the middle of the skill range. In the process, knowledge capital (design, patents, R&D, software, algorithms) finally overtook physical capital (machines, building tools) and human capital (education, skills, training). By the 1990s, knowledge owners (people who controlled knowledge capital) became the biggest beneficiaries of the technological system.

The latest technological transformation is by far the most radical, in terms of its potential to shape and alter societies and economies. In AI and ML, computers are ‘taught’ to ‘learn’ from huge sets of data. For example, a computer is taught to recognize patterns within massive amounts of data. For instance, a computer can be trained to identify a picture of a gun and identify it as a gun whenever it sees it. So, when this computer is shown a cover page of a detective novel, it will be able to identify the gun and tell us that there is a gun drawn on the page.

What the Digital Takeover Implies

The current digital turn is powered by the exponential growth of computer processing speeds (doubling every 18 months or so), driven by developments in the electronics industry. Concomitantly, data transmission rates are growing three times faster than computer power, driven by technologies such as fiber optic cables. Overall, massive computer networks are being built, creating huge computer-based systems, and encouraging greater digital networking and social media platforms.

A range of tasks are now being taken over by technology:  speech recognition, language translation, identifying diseases through X-rays and scans, typewriting, transcriptions, translations, to name a few. We even have robo-lawyer programs – software programs that will listen to a client and give legal advice based on loads of legal information fed into it. These will not displace the human lawyer completely, but it will surely displace a section of the traditional legal office.

Unlike previous technological transformations which primarily affected the manufacturing and agricultural sectors, this transformation is affecting the service sector which employs a large proportion of people today. The speed of this transformation certainly contributes to its overwhelming impact. World information flows have doubled every couple of years for decades. In contrast, world trade flows cannot possibly match this rate of growth.

The ongoing digital transformation is also reshaping competition. Human workers now have to compete with ‘digital workers’, another name for computer software. These digital workers form a new ‘ideal type’: they don’t get paid wages, they work 24x7, they don’t need any perks – holidays, lunch breaks, sick leave, they are not concerned about workplace regulations, they will never form a union.

Many jobs no longer need human beings to be physically present. Computers allow them to work from far away. So, companies can hire workers (from India, Bangladesh, Pakistan, Philippines) for much cheaper wages from across the globe.  These ‘tele migrants’ do a lot of work but have lower bargaining power and demand a lot less (lesser wages, no severance packages, no maternity/paternity benefits, no paid holidays or pension contributions). This is leading to a battle amongst workers for the lowest wages, and the worst working conditions.

Products are also becoming expensive. While there is a push towards cheaper digital components (large training data sets, computer servers storing data, internet) digital products themselves (made from processing/using digital components) are extremely costly and valuable. For example, self-driving cars are very expensive, though they are based on components such as GPS/Wi-fi/ sensors, anti-lock brakes which are cheap or even free.

This transformation is unlikely to see diminishing returns. Economists say that human capital (education, skills, training) as well as physical capital (machines, building tools) show diminishing returns. There is a limit to the increase in profits on investment one can get from increasing human and physical capital. But knowledge capital does not follow the rule of diminishing returns; the more one invests in it, the more the output per worker continues to increase.

Clearly, this digital turn requires the state to intervene, if we do not want a massive social, political, and economic upheaval and disruption. State-sponsored income support, retraining programs and tax-and-redistribute policies, are the needs of the hour.