Education must adapt to AI-induced changes

By September 19, 2016
Intelligence artificielle & éducation

Now that artificial intelligence and robotics are bringing about far-reaching changes in the labour market, education and vocational skills training will need to evolve if they are to help people get the best out of these innovations in their working lives.

Much ink has been spilled on the subject of how the jobs market is being impacted by artificial intelligence (AI) and robotics. The well-known study by economists Frey and Osborne published in 2013, which predicts that 47% of all currently existing jobs in the United States will come under threat over the next twenty years, is regularly brought out of the cupboard as a terrifying spectre. Other far more optimistic studies, based on longer time-frames, have delivered a riposte to this – largely unfounded – scaremongering, which has in fact been repeated many times over throughout our history.

However, the potential impact of AI on general education and vocational skills training – two means of preparing people for the labour market – is still being largely disregarded. The model whereby you learn during the first half of your life and spend the remaining years applying what you have learned in the world of work has held up pretty well. Nevertheless, the impact of AI on the labour market is going to require profound changes in our education system, so as to enable as many people as possible to benefit from the new approaches. So what changes is AI now bringing about and how can the education sector adapt to them?

The importance of being adaptable

As a recent study by the Economist Business Unit in on the impact of AI on the labour market points out, the surge in robotics is leading rapidly to skillset obsolescence, so that many of today’s workers will need to re-orient themselves and acquire new skills during their careers. The standard education model which has predominated since the Second World War, by which students specialise in a given field and perfect their mastery of that subject, is becoming increasingly inappropriate. Why spend many years acquiring skills that are likely to become obsolete in a few years’ time? James Bessen, an economist and lecturer at the Boston University School of Law who is the author of ‘Learning by doing: the real connection between innovation, wages and wealth’, argues that in the near future “the most important skill will be learning to learn.” Rather than being educated to be hyper-specialists, people should be trained to be multi-skilled, adaptable, and adept at the sort of improvisation that might well be necessary for the changes they will meet during their working lives.

However, many of the new jobs being created as a result of advanced AI development still require some very specific skills and, after all, you cannot become a data scientist overnight. It is therefore likely that working people will soon be facing a twin challenge: to show versatility and adaptability, and at the same time rapidly assimilate a range of complex skills. “I believe you have to find the right balance. People will continue to study medicine, law or economics, but they’ll also need a basic knowledge of computing,‟ argues Israeli-American Joel Mokyr, who is Professor of Economic History at Northwestern University in the United States. He predicts that “with the surge in the gig economy, people will have to be more flexible, open-minded, curious and able to demonstrate that they really can adapt. But at the same time, they will also have to be able to absorb a huge amount of information in a particular field, which means specialising. It’ll therefore be essential for people to know how to access information easily, and to be able to learn to unlearn”.

An antidote to the AI-induced revolution

Another direct consequence of this phenomenon is that the education system can no longer be restricted to serving young people but will need to get outside traditional classrooms and university lecture halls and accompany people throughout their entire lives. This message has already taken hold in some quarters, as witness the surge of Massive Open Online Courses (MOOCs), which can be accessed online free of charge, usually in the form of short videos, plus tools to enable students to discuss the subject with others. The first such initiatives were developed by AI experts: in 2011, Sebastian Thrun, Professor of Computer Science at Stanford University in the United States pointed the way when he made his course ‘Introduction to Artificial Intelligence’ – which he gave with his colleague Peter Norvig – directly accessible online. At the same time Andrew Ng, associate professor at Stanford, also offered free online access to his course on machine learning. Since then, Professors Thrun and Ng have both established startups providing online education. Thrun has set up Udacity and Ng has founded Coursera. In 2012, two prestigious US universities – Harvard and MIT – jointly launched the non-profit company edX, which specialises in developing MOOCs, with Anant Agarwal – formerly Director of CSAIL, MIT's Computer Science and Artificial Intelligence Laboratory – as CEO. Leading Artificial Intelligence experts thus seem well aware of the need to adapt education to the changes being brought about by AI and are seeking to promote their own solutions. Thrun says he founded Udacity “as an antidote to the revolution AI is fomenting.‟ Ng believes that researchers working with artificial intelligence ought to take some responsibility for finding solutions to the potential problems thrown up by their work and claims that Coursera is his own contribution to this process.

Ageing brains and AI

However, the need to go on learning throughout one’s life will not be problem-free. As we age, our brains tend to assimilate information less and less efficiently so it is hardly reasonable to expect older people to continue learning just as they did in their student days. On the other hand, with the recent increase in life expectancy, a growing number of people are now working to a more advanced age. In a report entitled ‘Intelligence Unleashed – an argument for AI in Education’, educational publishing specialist Pearson Education suggests that everyone ought to have a lifelong learning ‘companion’, a device equipped with ‘smart’ software to accompany, assist and encourage them throughout their studies. But if education is no longer restricted to our tender years, it will certainly need to be fully relevant to our working lives. So it is highly likely that what we generally call ‘education’ and ‘vocational skills training’ will become more interwoven in the future. “A significant part of people’s learning will no longer take place in the classroom but at work,” predicts Jim Bessens, adding: “Young technologies often change too quickly or are too unproven to be written into schoolbooks and taught in class. Meanwhile schools can’t keep up, they can’t keep on hiring new teaching staff who possess the latest skills, etc. So it’s not just a matter of acquiring new skills but of acquiring them informally, through experience and by communicating with other people.”

Sebastian Thrun’s Udacity offers ‘nanodegrees’ and micro-diplomas. These programmes can be completed in a few months so that people can continue to work while studying. Udacity also offers students a financial inducement to obtain their qualifications as quickly as possible: those who graduate in under twelve months receive a 50% refund on their tuition fees, which come to around $200 a month. Meanwhile the LinkedIn professional social network has set up its own online learning platform, The company sees promising synergies between the social network, its training arm and the users of both platforms. The LinkedIn algorithm could for example identify any gaps between a site user’s current skills and those required to obtain the sort of job s/he is looking for, and suggest suitable courses to help bridge the gap.

The lines between people’s education and working lives will become increasingly blurred

Strengthening the links between universities and companies

Jim Bessen thinks this ‘skills gap’ is going to be very hard to bridge. “Companies are already complaining that the institutions that have traditionally been responsible for educating future employees are no longer transmitting the knowledge required to take advantage of the new technologies. Firms are under-investing, fearful that their newly-trained staff will go elsewhere, and employees don’t make much of an effort anyway unless the skills they can acquire at that company are going to be useful to them elsewhere,” he explains. The Boston University economist reckons that “the most promising attempts to square this vicious circle include some new approaches, such as certification of skills acquired at work and collaboration between universities and companies so their employees are able to learn both in the classroom and in the workplace.‟ Siemens for instance has set up a partnership with the University of North Carolina (UNC) at Charlotte which allows employees of the wind turbine factory in Charlotte to study in parallel at UNC and obtain a diploma. So the Siemens staff gain on all counts: they acquire skills that will be useful on the job market and add to their CVs, without having to pay the astronomical college fees prevalent in the United States. So while changes due to AI may be proving a challenge to the education system, they can also provide an excellent opportunity to improve and explore new educational approaches.

Siemens is enabling its staff to obtain qualifications from UNC Charlotte

Towards the fluid corporation

As companies strive to adapt to these changes, they will, going forward, probably need to adjust the form in which they operate today. Predicts Joel Mokyr: “The corporation of the future will be a much looser association of various flexible, mostly small-scale, units. As 3D manufacturing and on-demand services proliferate, I think Uber- or TaskRabbit-like outfits are more plausible than General Electric. Remember, the big corporation is a modern invention – there were very few of them before 1870.” So companies are likely to have a more fluid structure, and workers will have to ensure their skills are up-to-date and keep growing their human capital. Which means that “we have to change the way we educate people; we must get the idea across to them that they need to continue learning. Everyone needs to look after his/her own human capital, just as you would maintain a house or a car,‟ he stresses.

Last but not least, Joel Mokyr foresees that, as automation takes over most routine jobs, leaving people to concentrate on activities which allow them to fully exploit their potential, working in these new-style companies is likely to be more rewarding. The Economic History Professor predicts: “In the fairly near future, we will basically get rid of boring routine jobs, and in the end only the people who want to work because their work is fulfilling and fun will work. Just as nobody sells subway tokens or sorts suitcases anymore, people will do things that, on the whole, are fun. I hope people will not just be playing Pokémon Go. But even if they do (or play some virtual-reality version of it), it beats selling subway tokens.”

Legal mentions © L’Atelier BNP Paribas