Rise of Artificial Intelligence, the Threat of a Jobless Future or Better Job Opportunities Through Reskilling and Upskilling

In 1811 groups of textile workers in the English Midlands began to smash the new weaving frames that they believed were taking their livelihoods. They took the name of a mythical leader, Ned Ludd, and the army had to be sent to suppress them. The Luddites were not wrong that machines destroyed their jobs. They were wrong, in the long run, that there would be no jobs left. Industrial employment grew for the next century and a half. The question posed by the topic is whether artificial intelligence will follow the same course, with loss for some and new opportunity for others, or whether it marks a break.

The pessimistic view has a long pedigree. In 1930 John Maynard Keynes wrote of technological unemployment, a new disease caused by our discovery of means of economising the use of labour outrunning the pace at which we can find new uses for it. In 2013 the Oxford researchers Carl Benedikt Frey and Michael Osborne estimated that about forty seven per cent of jobs in the United States were at high risk of automation within a couple of decades. Since 2022 systems that generate text, images and code have extended the reach of machines to work once thought to be uniquely human, such as writing, analysis and design.

Several features of AI justify concern. Previous waves of automation mainly replaced physical and routine tasks, and displaced workers could move into services and clerical jobs. AI affects cognitive tasks, which are the refuge of the educated. It can improve at great speed, and it can be copied at almost no cost. Brynjolfsson and McAfee argued in The Second Machine Age in 2014 that digital technologies may produce a world in which the gains go to those who own capital and rare skills, while the median worker's wages stagnate.

For India the stakes are high. About twelve million young people enter the labour force each year, and the country has relied on services, including information technology and business process outsourcing, which together employ several million workers, for much of its recent growth. These sectors are exposed to automation, since many routine office tasks, from data entry to call handling and basic coding, are those that AI does well. Agriculture still employs close to half the workforce at low productivity, and manufacturing has not created jobs on the scale of East Asia. A shock to services could slow the whole structure of Indian development.

A shock to services could slow the whole structure of Indian development.

Yet history gives grounds for a more hopeful view. Technology typically changes the composition of work and not the total. When automatic teller machines spread in the United States, many predicted the end of bank tellers. The economist James Bessen found that the number of tellers actually rose for decades, because cheaper branches meant more branches, and tellers moved from counting cash to selling services. The World Economic Forum estimated in 2018 that by 2022 about seventy five million jobs worldwide would be displaced by machines but one hundred and thirty three million new roles would emerge. Occupations such as data analyst, machine learning specialist and renewable energy technician barely existed a generation ago.

Economists also stress that automation often complements human work. A radiologist helped by software to read scans may handle more patients, a teacher with an adaptive tutoring programme may reach more students and a farmer with a weather prediction app may waste less. Jobs are bundles of tasks, and machines replace some tasks while leaving others, so that the person's role shifts toward judgement, empathy and creativity. David Autor has argued that the demand for human skill has repeatedly grown with technology, though unevenly.

The crux is reskilling and upskilling, as the topic recognises. Displaced workers need to acquire new capacities quickly, and the young need an education that prepares them for a changing market. Singapore's SkillsFuture programme, launched in 2015, gives every citizen a credit for lifelong training. India launched the Skill India mission in 2015 and has trained millions in short courses, though evaluations show that the quality and the match with jobs have been uneven. Germany's apprenticeship system links firms and schools. A country that treats learning as a lifelong activity will be better placed. Employers too must decide whether to treat their workers as costs to be cut or as capital to be developed.

But reskilling is no magic cure. It is hard for a fifty year old clerk to become a data scientist, and the retraining programmes of many countries have had modest success. Opportunities are unevenly distributed, with the best training offered to those who are already educated and urban. Women, rural workers and the informal sector, in which most Indian workers are employed, may be left behind. The transition may also be faster than the training system can match, leaving a period of dislocation whose human cost is real even if the long run is bright. Policy needs to include support during the transition, such as income support, portable benefits and help with mobility.

The foundation for the future is education that builds adaptable minds. Literacy, numeracy, reasoning, communication and the ability to learn, which machines have not replaced, are the lasting assets. Schools that drill facts which a machine can recall instantly prepare students for a world that has gone. Equally important are the ethics and governance of AI, such as questions of who owns the data, how algorithms treat people and how the gains are shared, since the distribution of benefits is a political choice and not a technical fate.

The topic poses a dichotomy between a jobless future and better opportunities, but the likelier outcome is a mixture, the proportions of which will depend on decisions yet to be made. Technology sets the possibilities, and societies decide which of them to realise. What no one can say is whether this time the machines will do what they have never done before, which is to learn faster than the humans who are to be retrained to work beside them, and the history that consoles us cannot say if it has an exception.