Are universities doomed to fail us?

At a time when society is accelerating, universities can't afford to produce important knowledge too slowly

Universities are transformative institutions. We could spend the rest of our lifetimes discussing the amazing inventions, innovations, and cultural movements that have blossomed as a result of these higher education bodies.

The work of Joseph Lister in the 19th Century at Glasgow University developed the use of antiseptic surgery. Manchester University oversaw vital development of the contraceptive pill. Coldplay met at UCL. Despite that final point, there is no question that universities have been essential institutions in catalyzing the progress of society by scientific, economic and cultural means.

But universities, in their current academic structure, suffer from a “lagging” problem. This has the potential to be a significant risk for an increasingly accelerating world facing multiple humanity-defining problems.

The “lagging problem” is as follows. Academics may become inspired by a subject whilst in their early twenties, perhaps as an undergraduate or on grad school programs.

This may set them on a particular intellectual path. Although you can switch between disciplines or sub-disciplines, committing to one set of knowledge problems can preclude you from developing the skills to deal with others. 

Because the route to tenure/seniority is now so long, it can often take at least 15 years from that initial moment of inspiration to leading teaching on courses/publishing widely 

By this point, many of the sources of inspiration that these professors specialised in may be of less pressing importance. However, these sources are also their bread and butter (and what they are monetarily and socially incentivised to continue researching), and therefore such topics are often what academics will teach to the next generation of students. As a result, current crops of undergraduates are being fed a weird diet of knowledge which won’t always equip them for the problems of the future.

To me, this partly explains this sinking feeling that many students get when they look at their module choices and say that everything seems a little bit outdated. I couldn’t study any modules on the social sciences of AI or technology policy, yet when it came to theories of justice, US foreign policy, or finance economics, there was an embarrassment of riches.

Nevertheless, this lag effect may always have existed in some form. Yet society, in the past hundred or so years, has done OK so far, right? So what is the problem here?

The hinge of history

The potential problem lies in how important you see this coming century for the future of humanity.

In the effective altruism community, of which I am a keen follower, there is an increasing emphasis on longtermism.

A philosophical school of thought committed to future generations, longtermists believe that the potential for humanity to oversee thousands of new generations, each with increasingly healthier and wealthier lives, means that greater efforts should be made to reduce the likelihood of global catastrophic events that could wipe out our species. Consequently, resources and time should be channelled into projects that span hundreds and thousands of years into the future.

There is, however, major debate within the EA community around what time scale we should put those resources into. The answer to this question, where one should focus their resources, depends on when you think the most important time in human history will be. 

Richard Fisher, who blogs at “The Long-termists Field Guide”, elegantly lays out the hinge of history hypothesis as applied to the next hundred years:

“According to a number of serious researchers, there is the chance that the 21st Century will see the arrival of sophisticated artificial general intelligence that could quickly evolve into a superintelligence. They argue that how we handle that transition could determine the entire future of civilisation, through a kind of “lock-in”.”

If you buy this logic, then the potential for AI to massively upscale growth and transform lives this century means that nailing things now is very important. The social and economic upheaval could be overwhelmingly positive, but if handled incorrectly, could either destroy humanity or significantly reduce its potential.

But what does this have to do with universities?

The consequence of accepting this point, that the next hundred years are the most important in our long term future, mean that academic lagging effects of 15 or so years are a major risk factor to humanity failing to secure a vibrant, flourishing footprint in the cosmos.

We need many more people to understand what AI can do now, but also what it will do in the future, and how we can manage the alignment of machine preferences/values to human values.

Claims of this kind are the subject of Azeem Azhar’s upcoming book (I will hopefully review this), where technological advancement outstrips the ability of society to catch up.

Pushback on these assertions are fair. There is a lot of fierce disagreement over the trajectory of AI - forecasting models are often very varied, with many experts not thinking that superintelligent AI could be achieved in the next hundred years .

But the same levels of humanity-defining urgency could be levelled at climate change, which has a lot more forecasting consensus regarding its development.

In the same sense, we do not have time to wait around for a couple of generations of academics to help carve out an alternative planetary management model. Time is of the essence.

So what are the alternatives?

In order to deal with these problems, one of the things we need to do is improve both universities and alternatives to universities.

One potential route is through rewarding younger researchers further through earlier professorships and publications.

In Nature, for example, the share of papers authored by “chaperoned” (i.e. published authors co-authoring with junior, unauthored academics) senior authors grew from 16% to 22% between 1990 and 2012, while new senior authors dropped from 39% to 31%. 

This isn’t just a problem in universities, but with the public and the private sector more broadly. Obviously, it does take time to build up deep domain knowledge (something that does favour lagging), but such steep structures inhibit the best ideas from young researchers being recognised. Ultimately, if we believe that certain younger people are going to have the ideas and enthusiasm around the right problem areas, their careers need to be accelerated.

Another avenue is to make it easier for startups to form using university research. In much of the UK and Europe, academic entrepreneurship is actively discouraged. Negotiations over intellectual property rights with technology transfer offices (TTOs), as well as demands of up to 50 per cent of equity shares in the potential spinout mean that generating these valuable companies often isn’t worth the hassle.

According to Air Street Capital partner Nathan Benaich, in Europe only 4 out of the top 116 VC backed unicorns are university spinouts. This needs to change.

In a piece for the FT, Benaich suggests we do the following:

• That we implement a globally competitive standardised deal for TTOs of either; equity share between 1 and 5%; 1% of global royalty; or 1% of global exit value upon M&A or IPO

• Improve alumni ecosystems that feed future innovation through donations and mentorship

• Universities should task TTOs with maximising the number of spinouts they have

The point on spinouts hints at my final suggestion, which is that we must create alternative institutions to universities. Ones that can be more nimble, or have higher capital available.

Each institution can have different, competing missions. After all, although we may have some confidence about what is important or will work on a particular problem, we do not have absolute certainty.

This can take many forms. Governments, for example, have much more capital at their disposal and can throw money at a problem if the institutional will is there.

Alternatively, the Fast Grant mechanism being developed by Patrick Collision and Tyler Cowen quickly turns around funding with less constraints on how the money is spent. With academics spending up to 25% of their time on writing funding proposals, more efficient ways of allocating resources could open up significantly more time to doing vital research.

When it’s all said and done, academic institutions are essential in our attempts to deal with the world’s most pressing problems. One of the issues with these kind of bodies is that they are often too slow, steeped in tradition and hierarchy. At the hinge of history, sometimes you need to move fast and fix things.

Of the Week - My Favourites

Podcast: New Statesman - Duncan Weldon on economic history, pandemics, and the politics of printing money

Youtube Video: Adult Swim - Joey Pera Puts You To Sleep

Song: KURUPT FM ft Chabuddy G - Aldona

Article: Matt Clancy - The “idea” of being an entrepreneur 

Matt is one of the sharpest minds working on innovation policy. In this post, he looks at the ways in which entrepreneurial activity can be passed from parents to kids, as well as in companies. The nuanced and complex gendered dimension here is really fascinating, and well worth your time.