‘Multiversities’ a key weapon in beating the bots

The university of the future will need to be a very different beast.

Illustration: Eric Lobbecke

What lies in store for future economic growth? It is difficult to know with any confidence. But if history is any guide, the story of growth will hinge on the interplay between the two i’s — the disruptive forces of innovation on the one hand, the stabilising role of institutions on the other.

There is an active debate under way on the first of these — the likely future path of innovation. Some have recently argued that the pace of innovation may be declining. The argument here is that much of the low-hanging fruit from the ICT revolution has already been picked. Diminishing returns to R & D may have set in. And other secular forces — including adverse demographics and rising inequality — may have acted as additional headwinds to growth. This view sometimes goes by the name secular stagnation.

An alternative school of thought believes we may instead be on the cusp of a new great wave of innovation, a fourth industrial revolution. This revolution in innovation will, it is said, be jet-propelled by a set of potentially new general-purpose technologies, among them AI, big data and ­robotics. For technology optimists, the story is one of secular innovation, not stagnation. And on this view, the potential gains in productivity and growth could be as large as in the case of any of the earlier industrial revolutions.

Typically, the two sides of this argument are taken to be in a secular struggle: the dark forces of secular stagnation pitted against the dynamic forces of secular innovation. Yet in practice both arguments have merit. They are certainly not mutually incompatible. It is perfectly possible to imagine a world of rapid innovation that nonetheless leaves large swathes of society in its slipstream.

Every industrial revolution has resulted in a hollowing-out, typically of mid-skilled tasks. Historically, that has meant largely manual, labour-intensive tasks. Machine has replaced human in activities that are routine and ­repetitive.

The future could well be very different. For example, the dawning of AI means that humans will no longer have the cognitive playing field to themselves. Thinking or non-routine tasks may increasingly be taken up by machines. They will be able to process more quickly, more cheaply and with fewer errors than their human counterparts, at least in some ­activities.

That could make the hollowing-out of human tasks, now cognitive as well as manual, far greater than before. A cottage industry has emerged over recent years manufacturing estimates of job losses that might result from this rise of the robots. As an early contributor to this industry, I am at least partly to blame. The estimates of gross job loss span a pretty wide range, but lie anywhere between 10 per cent and 50 per cent of the global workforce, depending on whether it is jobs or tasks that are assumed to be displaced.

Even at the lower end of this range, the societal impact would be significant. At the upper end, they would be truly transformative. If the truth lies in between, this could still make the job losses from the fourth industrial revolution greater than its predecessors. And if so, then the potential societal costs — such as rising wage inequality and threats to ­social cohesion — could also be as great as ever previously.

Some of the potentially powerful effects of automation on jobs and wages are already apparent. In the US, each industrial robot per thousand workers has been found to reduce the employment rate by 0.2-0.3 percentage points and wages by 0.25-0.5 per cent.

In Europe, results are more mixed. Research suggests that technology may have been the largest single contributor to falling labour shares in the recent past. Whether these effects on jobs and pay will be temporary or permanent is far harder to judge. At this stage of the technological cycle, the evidence is more likely to be picking up the shorter-term displacement effects of technology than their positive longer-term effects on pay, productivity and the demand for new goods and jobs.

It is the balance of these two effects, each large in gross terms, that will ultimately determine where unemployment will settle. No one could say with confidence today what the long-run net effect will be. What can be said with confidence is that the scale of gross displacement may be larger than in the past.

This means that if technological unemployment is to be avoided, positive effects on productivity, goods and jobs will also have to be commensurately larger. That has led some to argue that technological unemployment is more likely this time around.

Even if it is avoided, the societal costs of transition could well be larger, in terms of rising wage inequality, a falling labour share and damage to social cohesion.

And if that recessionary impact of technological change on individuals and societies is sufficiently large, it could well call into question the merits of having pursued the creative course in the first place. Or that, at least, is the lesson of history.

What experience since the Industrial Revolution has taught us is that this risk can be mitigated by an appropriate institutional response. To be effective in curbing the risk of recession, that response should have as its objectives, first, speeding up the process of reskilling by workers (“enabling”) and, second, cushioning the impact of new technologies on displaced companies and their workers ­(“insuring”).

In the face of mounting concern about the societal impact of the rise of the robots, quite a lot has been written recently on the sorts of new insurance mechanism that might be required to smooth this transition. Here, I will focus on two areas where new “enabling institutions” could ease the transition for workers and companies.

Both draw on one of the greatest institutional inventions of the past millennium — universities.

In 1852, John (later Cardinal) Newman published The Idea of a University, a book that has since assumed classic status as a statement of the principles by which universities should be run and organised. There are many themes to that text. From these I would highlight two that have particular relevance to the debate about the future role of universities within our societies.

First, Newman emphasised universities’ role in providing a liberal education for students, by which he meant access to a range of disciplines. Rigid specialism was to be avoided. The capacity to think was what mattered.

Second, the role of the university was to propagate ideas, rather than necessarily generate them. The capacity to diffuse knowledge, rather than generate it, was what mattered. With a twist, both are relevant to the “idea of a university” in the 21st century.

For decades, the primary focus of these institutions (universities) has been on providing young people with cognitive skills. That model has worked well in meeting the needs of the skill-shifts seen through each of the three industrial revolutions.

Whether that model will meet the needs of the fourth industrial revolution seems more questionable, for two reasons.

First, in future it seems very likely young and old alike will be in equal educational need — to train, retrain and retrain again through their 60 or 70-year careers. Second, the skills these people require for the future world of work will no longer be cognitive. Rather, they are likely to be more evenly balanced between the cognitive, technical and social — head, hands and heart.

It has become rather trite to talk of the need for “lifelong learning”. But never has the need for such learning likely to have been greater, given the longer span and greater volatility in future career paths. Making these career transitions will itself call for a particular set of skills — personal resilience, problem-solving and flexibility. As is now well-recognised, these character attributes are best instilled in early years.

If we now turn to our existing universities and colleges, at present these do not appear to be ideally sited to meeting either of these secular shifts in educational and training need. By and large, they are not currently configured as centres for lifelong learning. And nor, in the main, are they institutions providing a balanced educational offering of cognitive, technical and social skills.

To meet the needs of the future world of work, that might need to change. The importance of a broadbased, Newman-style education is likely to be greater than ever. This might crisscross disciplinary boundaries, as one way of increasing people’s ability to make giant logical leaps. Those skills will be social and technical every bit as much as cognitive, with head, hands and hearts sharing equal billing.

Put this way, the future university may need to be a very different creature than in the past. It may need to cater for multiple entry points along the age distribution, rather than focusing on the young. And it may need to cater for multiple entry points along the skills spectrum, rather than focusing on the cognitive. It would, in short, need to be plural rather than singular — a “multiversity”, rather than a university.

Andy Haldane is the chief economist of the Bank of England. This is an edited version of a speech he gave at the University of Oxford on May 23.