Existential risk is the possibility of humanity's long-term potential being destroyed. This includes both humanity going extinct and unrecoverable civilizational collapse. This could look like an engineered disease, or artificial intelligence, or something we haven't foreseen.

In the abstract, it can sound less menacing. In reality, this would mean everybody you love dying, everybody you know dying, all children dying, etc. This is a future worth avoiding! Furthermore, it means the loss of humanity's long-term potential. In the future, we could engineer the world to be better for all forms of life. We shouldn't miss our chance to try.

• You can work on a career in researching and mitigating existential risk.
• You can donate to organizations doing the above.
• You can pressure your representative to fund chronically underfunded x-risk insitutitions.

• Green branches are ones where humans do not suffer a catastrophe in the next 100 years (per this Metaculus question).
• Yellow branches are ones where humans suffer a catastrophe in the next 100 years.
• Red branches are ones where humans suffer a near-extinction event in the next 100 years.
• 'Catastrophes' are defined as events that reduce the human population by at least 10%.
• 'Near-extinction events' are defined as events that reduce the human population by at least 95%.

Yes and no. They are averages of predictions from the Ragnarok Question Series (developed by Tamay Besiroglu) on the forecasting site Metaculus. Popular predictions in the media are often unquantified. The Metaculus numbers are 'best guesses' from forecasters and are a step up over simultaneously certain and unspecified predictions.

However, the Metaculus predictions are probably biased to be optimistic, because forecasters can safely predict that humanity will survive: points won't matter if everybody dies. See this comment here for more.

Better numbers could come from domain experts. Even better numbers could come from 'adversarial collaboration' between general forecasters and domain experts. We are happy to include alternative datasets on the site from experts, or from famous predictions in our collection, which has started with data from Michael Aird's Database of existential risk estimates.

That's awesome! The creator of this site agrees. However, humanity dying out is so bad, that it's still worth working on reducing its chances. A 1% chance of billions of people dying is too high.

It can be! One thing that can help is to keep in mind the potential behind all the futures where we survive. A recent video from Kurzgesagt does a great job of illustrating that potential.

Conor Barnes made this with funding provided by the Long-Term Future Fund. Predictions are pulled from Metaculus.

Yes, right here.

The tree is based on Mike Bostock's Radial Cluster Tree layout (Copyright 2022 Observable, Inc. Released under the ISC license. https://observablehq.com/@d3/radial-cluster)

• 80,000 Hours (2020) Policy and research ideas to reduce existential risk.
• Existential Risk, Effective Altruism Forum - Topics.
• Bostrom, Nick (2002) Existential risks: analyzing human extinction scenarios and related hazards, Journal of Evolution and Technology, vol. 9.
• MichaelA (2020) Database of existential risk estimates , Effective Altruism Forum.