On the Nature of Technology

There is a great deal of debate these days about the likely impact of Artificial Intelligence (AI) on the future of jobs, skills and education. The general tone of discussion is that AI will benefit the highly skilled and will leave the less skilled without jobs. While it is true that a new technology will benefit those who are specifically skilled in that particular technology, the impression being given is that AI will impact all sectors and leave behind the unskilled and less educated. But, what is the evidence for this?

This article makes the case that the global debate on AI and other emerging technologies is derived from a fundamental misunderstanding of the nature of technology and its impact on human civilization. Let us start with a very fundamental question: what does a technology do? A technology either makes it easier to do something that was earlier difficult or it enables an activity that was earlier not possible. Think of any technology and it will fit this general framework.

Let us now look at the long history of technology and consider the implications of this framework. Take the example of hand-held firearms. Prior to the spread of guns, armies around the world primarily relied on archery for projectiles. When fire-arms first arrived at the end of the medieval period, they were less accurate, had less range and had a poor rate of fire. Indeed, it would not be till the Napoleonic Wars of the early nineteenth century that muskets significantly exceeded the range and accuracy of Mongol archery of the thirteenth century. It would be another half a century before they exceeded the rate of fire.

So, why did hand-held guns spread rapidly in the sixteenth centuries despite the fact that they were less efficient than the established technology? The answer is in the ease of use. It took many years of training to achieve proficiency as an archer or become a mounted-knight, but only a few weeks of practice for a peasant to fire a musket. Medieval armies usually had small contingents of trained archers and heavy cavalry – the rest were poorly trained infantry cobbled together from the general population. Any group that could gather together a larger group of trained mounted archers – such as the Mongols – had an overwhelming advantage. Guns reversed the equation by allowing a barely trained peasant to fight back. In turn, armies began to grow in size as mass-produced muskets (often of mediocre quality) allowed mass deployment of soldiers on a scale that would not have made sense earlier.

Who gained and lost from the shift? Obviously, those who were specifically proficient in the new technology (for example, gun-smiths) benefitted from the shift, even as specialized fletchers, who made arrows, became obsolete. However, the wider impact was to make recently conscripted and less skilled infantry-men outcompete highly trained archers and elite knights with expensive equipment. This democratization of military power ultimately ended the feudal structure. Over the centuries, gun technology evolved from the arquebus to the musket, then to the rifle, and eventually to today’s semi-automatic rifles. Modern rifles are obviously far better than archery but hand-held guns would never have got off the ground if they had waiting for skilled users.

A similar story can be told of the Industrial Revolution. Till the late eighteenth century, India was home to the world’s largest textile industry. This was a position that India had held since the Bronze Age. Indian weavers were highly skilled and often organized in large guilds for mass production. The wide variety of textiles and designs produced by them had been sought after by the Roman nobility as well as the French aristocracy during the reign of Louis XIV (Indian “chintz cotton” was such a craze that its import has to be controlled). Then came the Industrial Revolution at the end of the eighteenth century and machine-made textiles simply took over the world. The impact on India’s industrial base was so devastating that many parts of the country deurbanized, and the hand-turned charkha became a symbol of protest against the British.

In the initial phase, the quality of the machine-made cloth was inferior to that produced by Indian weaver guilds, but the quantity of production was on a totally different level. Eventually, the quality of machine-made cloth would catch-up and surpass handloom for most forms of textile. The Industrial Revolution benefitted the British engineers who designed the steam-powered machines, and the mill-owners who deployed them. However, from a job creation perspective, it allowed unskilled English peasants, newly evicted from traditional farming by the Enclosure Movement, to compete with highly skilled Indian weavers. Indeed, one of the factors that allowed the Industrial Revolution to succeed was the fact that low-skill industrial workers, with only a working knowledge of one segment of the production chain, could take on the accumulated knowledge of high-skill guilds that had functioned since medieval times.

Readers can see from the above examples, that the wider impact of technology is to allow the unskilled to compete with the skilled. The accumulation of technology and social organization allows us today to get a meal online (a low skill activity) rather than having to hunt a woolly mammoth (a high skill activity). The owners of online food delivery platforms obviously benefit from the technology, but the average consumer also avoids the risk of being chased by an angry herd of woolly mammoths.

This brings us to what to expect from Artificial Intelligence and other emerging technologies. The most visible beneficiaries will be those who invent and control the technology, but the wider impact will be to allow low-skilled workers to compete with high-skill workers. The average citizen will be able to draw up a tenancy contract without a lawyer, an amateur script-writer will be able to compete with a famous movie director, a newbie engineer will be able to print out a machine-part that would have required advanced forging skills. This is democratization, not mass unemployment. The key is to absorb and deploy the new technology, rather than some misplaced Luddite attempt to “protect” old jobs.

Not all professions will experience the above disruption but, for many areas, artificial intelligence will allow the relatively unskilled (or uneducated) to provide services at par with far more skilled and educated workers. A few elite professionals will probably not be impacted (and may be even benefit from AI) but the real change will be massification of many high-skill professions. This is not doomsday for all jobs. Certain expensive services and goods will become cheap and commonly available. In turn, this will benefit consumers and businesses that buy these products. Completely new professions will emerge. With many previously expensive services and goods being available cheap, new multi-disciplinary combinations will emerge: hospitals designed by medical professionals rather than traditional architects, and bespoke fashionable furniture designed by your local carpenter.

Of course, there will be pain during the transition, but policy response must be flexible labour markets and platforms for re-skilling (not necessarily up-skilling as low-skills may be adequate). An important implication of such a world of accelerated innovation is that young people should enter the workplace early rather than spend too many years becoming ever more trained in skills that could soon become outmoded. Instead, tertiary education should be something that is delivered throughout a person’s life – apprenticeships, a certificate here, an online course there, and the likelihood that most people will have multiple careers in the course of their lives. This is iterative updating of skills rather than a unidirectional waterfall.

(Sanyal is Member, Economic Advisory Council to Prime Minister. Opinions are personal)