Decision-making in the face of global catastrophic biological risks

Are we prepared to prevent a global catastrophe? Are we doing enough to mitigate global catastrophic risks such as nuclear accidents, deadly pandemics, transformative artificial intelligence or extreme climate change? Are decision-makers well-equipped to prevent such low probability but high impact events?

On 6 September 2021, the Geneva Science-Policy Interface (GSPI), the Simon Institute for Longterm Governance (SI) and the Centre for the Study of Existential Risks at the University of Cambridge (CSER) convened international civil servants and diplomats to explore ‘Long-term decision-making in the face of extreme biological risks’. This post summarizes the official side event to the 2021 Meeting of Experts of the Biological Weapons Convention (BWC), hosted at the Domaine de Penthes in Geneva.

The BWC was adopted in the 1970s, during the Cold War. The COVID-19 pandemic has accelerated discussions to adapt it to the modern age by providing a concrete example of the lack of coordination by the international community. See here for a more detailed account of recent discussions around the BWC.

But what are global catastrophic biological risks in the first place?

COVID-19 has given us a sense of global risks to our health and economy. But this pandemic does not currently fall in the category of global catastrophes. Global catastrophic risks are commonly defined as disasters that result in a loss of 10% or more of human life. To date, COVID-19 has caused around 5’000’000 deaths – way too many already but in absolute terms 0.0625% of the global population.

Dr. Lalitha Sundaram, researcher at CSER, qualified global catastrophic risks (GCRs), along three dimensions: the critical systems affected (e.g. health systems, economic systems, education systems, etc.), global spread mechanisms (e.g. natural global scale or via human activity), and prevention and mitigation failures (e.g. institutional, interpersonal or individual) .

The nature of such risks can be understood in various ways: some risks have existed since the dawn of humanity (e.g. asteroid impacts or supervolcanic eruptions), some risks are amplified by the growth of human presence and activity (e.g. ecosystem collapse, natural pandemics), others relate to technological developments (e.g. rogue AI, nuclear winter).

Dr. Sundaram pointed out that GCRs are not just about “short sharp shocks”. GCRs can also result from longer term trends such as the gradual reduction of fertility in human populations over many years. To design future scenarios of global catastrophic biological events, CSER leverages the Participatory Evolution methodology.

The figure below illustrates how global catastrophes differ from other systemic shocks: the disruption overwhelms the entire system such that it does not bounce back swiftly. Some global catastrophic risks can allow for recovery. Others are likely to eradicate life as we know it – so called existential risks.

From Stauffer et al., 2021

How to govern and make decisions in the face of low-probability high-impact risks?

Making decisions to prevent risks from happening involves processing a lot of complex information. Maxime Stauffer, Science-policy Officer at the GSPI and CEO at SI, explains how our ability to make good decisions is shaped by our behavior. 

The brains of Homo sapiens have not evolved to make such complex decisions. We are better equipped to deal with phenomena happening on shorter scales of time and space, such as extinguishing a fire we see in front of us. We can act swiftly in the face of immediate danger, but have more difficulty adapting to environments with many uncertain or rapidly evolving variables.

This is in part due to the fact that we rely on thinking by analogy (e.g. relying on past experience/history to deal with new events), are scope insensitive (e.g. if a problem doubles in scope, we don’t care twice as much) and struggle making our intuition explicit (e.g. defining terms, quantifying hunches), among many other reasons. Fortunately, there are solutions to build our capacity to make decisions in the face of global catastrophic risks. For example: 

  • Analogical training: The future includes many scenarios that we haven’t experienced. Analogical training is the process of simulating rare events such that participants can learn without their occurrence. Simulating the broad set of possible futures can prepare us to manage the unknown.
  • Group decision-making training: Training our capacity to make decisions as a group means improving how people exchange information, come to agreements, etc. Using the right processes to design and guide groups can reduce individual biases while leveraging each member’s unique point of view.
  • Training meta-cognition: In the process of making decisions under uncertainty, we have to get into the habit of observing and questioning our own thinking patterns. It is easy to underestimate extreme risks or forget about future generations. We need to step back and ask questions like: What are my goals here? How did I judge information? Humans are not defined by their thoughts and can improve their own thinking.

From theory to practice

The event ended with a practical exercise. The audience was tasked with prioritizing problems and improving pandemic preparedness plans. The simple task of filling out a survey already enables the observation of our own thought patterns. The two exercises, using pre-mortem and simple MCDA techniques, were a preview of the content of our training programs.

In November this year, on the side of the BWC’s Meeting of State Parties, the GSPI and SI will present a tabletop exercise on preparedness to engineered pandemics. Via role-play, participants will be immersed in decision-making scenarios. Such exploration applies the behavioral insights described above to help overcome some of the cognitive limitations we are all subject to.

Ultimately, the goal is to equip policymakers with the awareness and skills to build the future for generations to come. Of course, this task does not solely rely on our behavior, but it plays an important part.

If you are interested in participating, do not hesitate to contact us at maxime.stauffer @ unige.ch.