The term “existential threat” is used quite broadly in media and elsewhere.

Quite often, I think the term is used when it’s not appropriate.

This is a typology of how I think about different types of “existential” threat, in the way they are referred to. I’ve ordered them from what I consider to be really bad to really, really bad. (Note that I’m focusing on severity rather than probability.)

  1. Existential threats like climate change. It will, or could, make life very difficult, and the extreme worst-case scenarios could veer towards category 3 below (although I don’t believe this is the consensus view). Climate change is likely to have a number of second-, third-, and n-th order effects that will probably be negative, on net. It might also contribute towards increasing the probability of some of the risks below (as do all of the others; this typology is an over-simplification and doesn’t focus on how these and other threats can interact with each other and compound). However, it’s often referred to as being “existential” when I don’t think that’s actually the case.
  2. Existential threats that fundamentally change the way we live our lives, like shifts towards authoritarian governance that restricts our individual and group freedoms. Whether these come from the political “left” or “right” (whatever those terms mean to you), if freedoms of expression, movement, association, and pursuit of personal values are systematically removed, this results in an entirely different experience of life. It’s a stretch to say that this threat is existential, but it definitely has the opportunity to change the way we exist in an important way. I also think that this type of threat makes the other forms of existential threat more likely.
  3. Existential threats that don ‘t wipe humanity but puts technological or social progress back hundreds or thousands of years. I put the threat of nuclear war in this category, in a similar way to this sentiment, often attributed to Einstein: “I know not with what weapons World War III will be fought, but World War IV will be fought with sticks and stones”.
  4. Existential threats that wipe out humanity and its future entirely. This is why the idea of engineered biological weapons and unaligned artificial intelligence (in the “paperclip maximiser” sense) scare me.

To be clear: I think all of these are bad. I understand why people use the term “existential” but I also think that it’s a stretch to stay that some of these threats are existential.

In all cases, I believe we should put resources towards understanding the risks and managing them appropriately. This includes trying to avoid these risks or the worst-case scenarios associated with these risks, and also looking to mitigate and reduce them. I also note that while severity is important, probability is also relevant and this justifies and does not detract from spending resources on risks that aren’t as severe on this threshold (or that don’t even fit onto this particular spectrum).