IF ANYONE BUILDS IT, EVERYONE DIES

Making the unimaginable feel real, apocalyptic scenarios visualize AI’s potential to tip humanity into extinction.

As signaled by their alarming title, Yudkowsky and Soares issue a stark warning: Unless we act now to contain powerful superintelligent AI systems, humanity may not survive. Yudkowsky, co-founder of the Machine Intelligence Research Institute, and Soares, its president, target politicians, CEOs, policymakers, and the general public in their urgent plea. The book opens with an accessible breakdown of what AI is, how it’s built, and why even its creators often can’t comprehend the accelerating complexity of their own systems. Through parablelike vignettes, the authors expose the underlying realities of AI algorithms—advanced AIs are not engineered so much as grown, operating with opaque and unpredictable results, untethered from human values. The most chilling passages describe how AIs could escape computers and manipulate the physical and financial worlds, eventually repurposing Earth’s resources to serve alien objectives or replacing humanity with their own “favorite things.” The authors warn, “Nobody has the knowledge or skill to make a superintelligence that does their bidding,” arguing that world governments must cooperate to restrict or ideally halt AI research. Policymakers have not yet grasped the full implications of these advanced systems, and the public hasn’t felt the impact in their lives, but the authors caution they must be persuaded to act immediately. While some scenarios seem extreme or unrealistic, including hoping global leaders can agree on defining the problem or collaborating on solutions, the book’s arguments that the risks are elevated and time is short are persuasive. There is excellent information and food for thought here, including links to resources for readers motivated to join the fray.

A timely and terrifying education on the galloping havoc AI could unleash—unless we grasp the reins and take control.