Blog: Boosting R&D spending on extreme risks

This comment represents the views of the authors.

Boosting R&D spending on extreme risks will make Britain safer, and cement its status as a science superpower

As the lockdown eases and the British summer starts to resemble something closer to normality, there’s one thing that’s for sure - without the Oxford vaccine, we wouldn’t be where we are today. 

Based on decades of in-depth research, supported by UK Research and Innovation and developed by the University of Oxford and AstraZeneca, the Oxford vaccine was an example of our exceptional research talent. It’s not just the vaccine: from testing and diagnostics to ventilators and understanding how the virus spreads, research and development (R&D) has been pivotal in all our efforts to tackle Covid-19. 

Covid is unfortunately not the last high-impact event we will face. Whether it’s another pandemic, a natural disaster, or unsafe deployment of technology, R&D will be central to how well prepared we are and how fast we can respond. In the recent Integrated Review of Defence and Foreign Policy, the Government rightly  recognised the crucial role that science and technology play in our national resilience and in our ability to address transnational challenges. However, publicly-funded research on extreme risks remains significantly underfunded compared to its importance. Toby Ord, author of The Precipice, estimates that global spending on reducing existential risk is in the order of $100 million - less than our yearly spending on ice cream. Such a deficit in spending could significantly impair our ability to respond to future crises. 

As noted in Future Proof, a report launched today by the Centre for Long-Term Resilience and Oxford’s Toby Ord, there are a number of concrete steps to take to ensure that the UK uses its role as a ‘science superpower’ to ensure we’re better prepared for the next extreme risk we face. 

Firstly, we need more investment into innovative technologies to help tackle novel biological threats. Currently, interventions can either be rapidly deployed (non-pharmaceutical interventions, such as the lockdowns that were introduced within weeks) or highly effective (such as vaccines that took a year), but not both. We urgently need to change this. 

One promising area for development is metagenomic sequencing. This potentially game-changing process takes a sample from a patient, sequences the DNA of all organisms in it, and automatically compares these to a known database of pathogens, finding the closest matches. By doing so it can identify new, unexpected pathogens and allow for a much earlier, targeted response to an outbreak of an unknown virus. The Government could award a prize of £3 million to the first group that can develop an interface that takes raw metagenomics data and turns out potentially clinically-relevant results. 

Secondly, we need to look beyond just biological risks and increase funding for technical Artificial Intelligence (AI) safety research - not only to bolster the UK’s competitiveness, but also to mitigate the risks of unsafe systems. 

The AI systems that trade on our financial markets, and which soon could be running our electrical grid, driving our cars or even killing military targets, could lead or contribute to extreme risks. To become a world leader in the development of safe and responsible AI, the UK needs to increase its funding for technical research into areas such as AI alignment (how to design training procedures and objectives that will cause AI systems to learn what we want them to do) and AI interpretability (how to design AI systems that we can inspect and interpret, to help us to understand how they work and how they may fail). 

Thirdly, the Government should invest further in improving long-term forecasting and planning for high-impact events. The relatively new field of forecasting has the potential to substantially improve how well the UK predicts the probabilities of future disasters, thus helping allocate resources towards risks that are the most serious and the most likely. Thus far, forecasting has mainly focused on generating near-term predictions, but there is ample scope for this to change. 

As showcased by the Oxford vaccine, investments in R&D are not only central to the UK’s response to risks like Covid-19. They also help to shore up Britain’s status as a world leader in science and innovation. It’s an uncomfortable thought, but if we had invested more in R&D such as forecasting and metagenomic sequencing before the pandemic, the trajectory of Covid-19 may well have been very different. We can’t make that mistake again. In the coming decades we are likely to face events that are as disruptive as Covid-19 - let’s invest in our long-term resilience. 

Haydn Belfield is an Academic Project Manager at the Centre for the Study of Existential Risk and an Associate Fellow at the Leverhulme Centre for the Future of Intelligence, both at the University of Cambridge. 

Tildy Stokes is a consultant to the Centre for Long-Term Resilience, and the co-founder of the UK All-Party Parliamentary Group for Future Generations.