Sandeep Gopalan :
The question is not purely rhetorical. Indeed, research confirms the inexorable march of machines into the workplace. Early research by MIT professor David Autor analysed the potential for offshoring job functions and posited that routine and repetitive jobs were easiest to transfer offshore. This theory held sway for about a decade until Oxford’s Carl Frey and Michael Osborne identified the potential for occupations to be automated by assessing tasks performed by 702 US occupations and concluded that 47 per cent of occupations could be automated by 2030. Other research has come to similar conclusions, like for example the McKinsey Global Institute’s report in December 2017 finding 50 per cent of paid work activities are technically automatable based on demonstrated technologies. McKinsey also showed that third of constituent activities in 60 per cent of occupations can be automated, and that 3.2 million FTEs of Australian jobs could be displaced by automation by 2030. In Australia, the Office of the Chief Economist applied the Frey & Osborne model and determined that 43.9 per cent of employment has ‘high automation susceptibility.’ The OECD issued a report in March 2018 showing that 14 per cent of all jobs in 32 countries are highly automatable and another 32 per cent of all jobs are likely to undergo significant change.
To be sure, just because a task could be done by a machine does not mean that it will be. Regulation and other barriers may inhibit technology adoption and preserve jobs. Equally, machines may perform low level functions leaving others to humans, and new jobs may be created. Some point to the fact that ATMs did not destroy bank cashier jobs – in fact more people were hired as bank branches increased.
But this time may be different. The ability of machines to process vast amounts of data and the revolutionary growth in machine learning (AI) means that the functions solely capable of being done by humans are rapidly shrinking. Moreover, machines are tireless, work all hours, don’t demand rights (yet), can’t forget, and reduce errors.
Due to these advantages, there is a race to perfect the technology and make it available to a wider array of functions. For instance, recently IBM unveiled Project Debater, which is capable of debating humans. In tests, the machine responds to arguments with a better array of data and evidence than humans. It is also able to discern the more salient aspects of an argument and make self-deprecating jokes. Humans obviously are superior at bringing life experience into the debate. But the gap is shrinking and professionals such as lawyers may be at risk.
Google recently demonstrated its Assistant making an appointment with a hairdresser. Most observers could not tell that it was not a real person. Similarly, Amazon has Alexa, and an array of voice assistants powered by AI are likely to replicate secretarial functions efficiently.
If that’s not scary enough, Andrew Alliance, a Swiss company, has developed a liquid-handling robot that can serve as a lab assistant. Other robots assemble IKEA furniture, serve as retail assistants, and perform a variety of jobs once thought possible only by humans. Even surgery, medical diagnoses, and journalism are not safe from machines. Robots already perform or assist in complex surgeries, and IBM Watson has been helpful in making cancer treatment plans. Drones may soon deliver post and pizza, and self-driving trucks will hit the roads in the next 10 years.
Amidst these developments, the Indian university sector has been slow to respond. Degree programmes vary only slightly from their avatars many decades ago, and education is largely based on rote knowledge acquisition. This is unlikely to prepare students to defend their jobs against robots or to work with and alongside machines to solve problems.
Some might cavil that a university education is not aimed solely at producing workers. But that is unlikely to be persuasive. As the cost of education rises, both government funders and students are demanding returns for their investment in universities. Others argue there is no need for robots in India due to cheap labour. This is nonsense – if the labour is not fit for purpose, it will be replaced by machines. Therefore, it is both naïve to ignore market developments and irresponsible to refrain from optimising degree programmes for the future of work.
Let’s consider what that might look like. Evidence indicates that humanistic characteristics are valued whereas routine functions will be automated. Then, communication, problem-solving, empathy and service become highly important because humans enjoy advantages over machines in these areas. Unsurprisingly, examining university programmes and their outcomes shows inadequate attention to these aspects. For instance, employers report that Indian engineering graduates possess poor communication and language skills. A report by Aspiring Minds claims that 40 per cent of engineering graduates cannot comprehend English. Reading judgments issued by courts shows that English language skills have not been sufficiently taught by law schools.
The provision of these skills alone is insufficient. Indian university degrees are extremely siloed and rarely integrate interdisciplinary approaches to problem-solving. IT programs don’t systematically expose students to domain knowledge from other disciplines. This deprives students of knowledge about how tech solutions might impact their end-users and an opportunity to leverage IT to solve disparate problems. For example, consider social work students being equipped with technology education. Given the growing penetration of mobile phones, an array of social services could be delivered via these devices if students are educated to become problem-solvers. Similarly, if policing expanded use of IT, facial recognition and other tools could be better deployed to tackle crime.
At present, there is no cross-pollination of domain expertise with tech tools within the university curriculum. Solutions to social problems are left to the vagaries of the market thereby depriving the country of potential innovation from its sharpest minds in universities. Unsurprisingly, the number of patent applications filed and commercialisation of research by Indian universities is trivial.
These changes are not particularly costly. They require universities to better conceptualise their vision as key parts of the innovation and social problem-solving landscape. It requires recognition that a university is not just a diploma mill indifferent to employment or social outcomes. And that education is about applying knowledge perfected through experiential opportunities for students working alongside peers from different disciplines.
If India’s university sector does not adapt to the automation of work, the country’s massive demographic dividend will be wasted. Millions of young men and women will be ill-prepared for the demands of employment and social unrest may be inevitable. India’s universities must rise to the challenge by creating more multi-disciplinary programmes, embedding skills, and expanding experiential learning to create problem-solvers not just degree-holders.
The question is not purely rhetorical. Indeed, research confirms the inexorable march of machines into the workplace. Early research by MIT professor David Autor analysed the potential for offshoring job functions and posited that routine and repetitive jobs were easiest to transfer offshore. This theory held sway for about a decade until Oxford’s Carl Frey and Michael Osborne identified the potential for occupations to be automated by assessing tasks performed by 702 US occupations and concluded that 47 per cent of occupations could be automated by 2030. Other research has come to similar conclusions, like for example the McKinsey Global Institute’s report in December 2017 finding 50 per cent of paid work activities are technically automatable based on demonstrated technologies. McKinsey also showed that third of constituent activities in 60 per cent of occupations can be automated, and that 3.2 million FTEs of Australian jobs could be displaced by automation by 2030. In Australia, the Office of the Chief Economist applied the Frey & Osborne model and determined that 43.9 per cent of employment has ‘high automation susceptibility.’ The OECD issued a report in March 2018 showing that 14 per cent of all jobs in 32 countries are highly automatable and another 32 per cent of all jobs are likely to undergo significant change.
To be sure, just because a task could be done by a machine does not mean that it will be. Regulation and other barriers may inhibit technology adoption and preserve jobs. Equally, machines may perform low level functions leaving others to humans, and new jobs may be created. Some point to the fact that ATMs did not destroy bank cashier jobs – in fact more people were hired as bank branches increased.
But this time may be different. The ability of machines to process vast amounts of data and the revolutionary growth in machine learning (AI) means that the functions solely capable of being done by humans are rapidly shrinking. Moreover, machines are tireless, work all hours, don’t demand rights (yet), can’t forget, and reduce errors.
Due to these advantages, there is a race to perfect the technology and make it available to a wider array of functions. For instance, recently IBM unveiled Project Debater, which is capable of debating humans. In tests, the machine responds to arguments with a better array of data and evidence than humans. It is also able to discern the more salient aspects of an argument and make self-deprecating jokes. Humans obviously are superior at bringing life experience into the debate. But the gap is shrinking and professionals such as lawyers may be at risk.
Google recently demonstrated its Assistant making an appointment with a hairdresser. Most observers could not tell that it was not a real person. Similarly, Amazon has Alexa, and an array of voice assistants powered by AI are likely to replicate secretarial functions efficiently.
If that’s not scary enough, Andrew Alliance, a Swiss company, has developed a liquid-handling robot that can serve as a lab assistant. Other robots assemble IKEA furniture, serve as retail assistants, and perform a variety of jobs once thought possible only by humans. Even surgery, medical diagnoses, and journalism are not safe from machines. Robots already perform or assist in complex surgeries, and IBM Watson has been helpful in making cancer treatment plans. Drones may soon deliver post and pizza, and self-driving trucks will hit the roads in the next 10 years.
Amidst these developments, the Indian university sector has been slow to respond. Degree programmes vary only slightly from their avatars many decades ago, and education is largely based on rote knowledge acquisition. This is unlikely to prepare students to defend their jobs against robots or to work with and alongside machines to solve problems.
Some might cavil that a university education is not aimed solely at producing workers. But that is unlikely to be persuasive. As the cost of education rises, both government funders and students are demanding returns for their investment in universities. Others argue there is no need for robots in India due to cheap labour. This is nonsense – if the labour is not fit for purpose, it will be replaced by machines. Therefore, it is both naïve to ignore market developments and irresponsible to refrain from optimising degree programmes for the future of work.
Let’s consider what that might look like. Evidence indicates that humanistic characteristics are valued whereas routine functions will be automated. Then, communication, problem-solving, empathy and service become highly important because humans enjoy advantages over machines in these areas. Unsurprisingly, examining university programmes and their outcomes shows inadequate attention to these aspects. For instance, employers report that Indian engineering graduates possess poor communication and language skills. A report by Aspiring Minds claims that 40 per cent of engineering graduates cannot comprehend English. Reading judgments issued by courts shows that English language skills have not been sufficiently taught by law schools.
The provision of these skills alone is insufficient. Indian university degrees are extremely siloed and rarely integrate interdisciplinary approaches to problem-solving. IT programs don’t systematically expose students to domain knowledge from other disciplines. This deprives students of knowledge about how tech solutions might impact their end-users and an opportunity to leverage IT to solve disparate problems. For example, consider social work students being equipped with technology education. Given the growing penetration of mobile phones, an array of social services could be delivered via these devices if students are educated to become problem-solvers. Similarly, if policing expanded use of IT, facial recognition and other tools could be better deployed to tackle crime.
At present, there is no cross-pollination of domain expertise with tech tools within the university curriculum. Solutions to social problems are left to the vagaries of the market thereby depriving the country of potential innovation from its sharpest minds in universities. Unsurprisingly, the number of patent applications filed and commercialisation of research by Indian universities is trivial.
These changes are not particularly costly. They require universities to better conceptualise their vision as key parts of the innovation and social problem-solving landscape. It requires recognition that a university is not just a diploma mill indifferent to employment or social outcomes. And that education is about applying knowledge perfected through experiential opportunities for students working alongside peers from different disciplines.
If India’s university sector does not adapt to the automation of work, the country’s massive demographic dividend will be wasted. Millions of young men and women will be ill-prepared for the demands of employment and social unrest may be inevitable. India’s universities must rise to the challenge by creating more multi-disciplinary programmes, embedding skills, and expanding experiential learning to create problem-solvers not just degree-holders.
(Sandeep Gopalan is the pro vice-chancellor for academic innovation at Deakin University)
