If Anyone Builds It, Everyone Dies: Why Superintelligent AI Will Destroy Us All

Why might an artificial superintelligence lead to the extinction of humanity? And is there any chance of preventing it?

In 2023, hundreds of leading AI experts signed an open letter warning that human civilization faces a real and extremely serious danger. Yet governments and corporations continue the race to build systems that outperform people intellectually, while the world is almost unprepared for the consequences of such a breakthrough.

For several decades, two signatories of that letter—Eliezer Yudkowsky and Nate Soares—have studied how superintelligence might reason and act. Their conclusions are alarming: a clash of interests between advanced AI and humans is practically inevitable, and in such a confrontation we have no chance of winning. For what reasons and by what means could a superintelligence wipe out the human species? What needs to be done to survive? The authors clearly, thoroughly, and vividly explain their position and urge everyone to think about the existential threat hanging over all of us—while there is still time.