How likely is human extinction due to a natural pandemic?

A pandemic is the rapid spread of an infectious disease across a large region. Global pandemics will likely occur in the future, but their danger is very hard to estimate. So far, pandemics have received far more attention than other natural existential threats and our analysis below indicates that this level of attention is appropriate. A natural pandemic killing all humans is unlikely but plausible.

Today, a disease can travel faster than ever before. Thanks to long distance air travel, an infectious agent can quickly spread to all continents. In fact, most humans are now connected by one or more of our rapid transportation networks.

On the other hand, enormous effort is being expended to prevent pandemics today. Serious work can be done on many fronts simultaneously because pandemics arise out of interactions between large numbers of people, animals, and institutions. Unfortunately, this complexity also means that for the foreseeable future our defenses will not be perfect and our uncertainty will be extreme.¹

The complexity of pandemics is also the reason why straightforward extrapolations from history tend to be deeply flawed. Human civilization today is unprecedented in several relevant ways. We can’t study past examples of globe-spanning civilizations who actively tried to protect themselves from this danger. Knowledge of past pandemics is still incredibly valuable, it’s just very hard to meaningfully generalize to our situation.

Lastly, the vast majority of global pandemic threats are not existential risks. Obviously no disease so far has had the right set of characteristics to cause human extinction. More reassuringly, there are good reasons to believe that there is an evolutionary tendency away from “perfect” deadliness. Arguments such as the “trade-off hypothesis” indicate that extremely deadly diseases would be under evolutionary pressure to become less deadly in exchange for higher transmissibility.² That is, diseases can’t simultaneously maximize their deadliness and transmissibility.³

Unsurprisingly, estimates of this risk vary wildly; even expert commentators tend to be highly uncertain about the odds. Predictions also tend to lump together several different kinds of pandemics and thus can make the future sound much more dangerous than it really is. For example, a concerned expert may estimate that there is 50% chance of a global pandemic during the 21st century. Taken at face value this is extremely frightening, but with further scrutiny it becomes clear that the vast majority of that probability would be diseases with low death rates and certainly not pandemics that have a chance of killing all of humanity.

Furthermore, most expert risk estimates also include pandemics that are deliberately caused by humans (through viral engineering). In this sense, pandemics are similar to climate change: our advancing technology means that we can now cause catastrophes of these kinds. In both cases, human action is much more dangerous than what nature can throw at us. Human-caused catastrophes are a real danger that should be taken very seriously, but for the present discussion we are focusing on natural existential risks.

Rolling several publicized predictions together with what we’ve learned, we arrive at this estimate of extinction risk: 1 in 100,000 chance per year (or 1 in 1000 per century). This estimate places natural pandemics as the single greatest risk among the natural existential risks we consider to be the most dangerous.

1. For example, the US Centers for Disease Control and Prevention (CDC) assess influenza pandemic risk using ten distinct factors pertaining not just to properties of the virus, but also to attributes of the affected human and animal populations. CDC, Influenza Risk Assessment Tool, Accessed 2016-03-13. [↩]
2. Lenski, R. E., & May, R. M. (1994). The evolution of virulence in parasites and pathogens: reconciliation between two competing hypotheses. Journal of theoretical biology, 169(3), 253-265. [↩]
3. Alizon, S., Hurford, A., Mideo, N., & Van Baalen, M. (2009). Virulence evolution and the trade‐off hypothesis: history, current state of affairs and the future. Journal of evolutionary biology, 22(2), 245-259. [↩]