Former Google CEO Eric Schmidt, along with Dan Hendrycks and Alexandr Wang, has co-authored a critical paper titled "Superintelligence Strategy," which cautions the United States against embarking on an aggressive AI development initiative akin to the Manhattan Project. The authors argue that such a strategy could provoke international tensions, particularly with rival nations like China that would likely not remain passive. They draw parallels between the current AI landscape and the historical nuclear arms race, suggesting that a rush to develop superintelligent AI could lead to destabilization rather than security. Instead of advancing aggressively, the authors advocate for a more defensive approach, recommending tactics such as sabotaging rival AI projects and ensuring domestic manufacturing of crucial AI components. Their threefold strategy aims to balance competitiveness with strategic stability, emphasizing that a careful, methodical approach could harness AI's benefits while mitigating risks of catastrophic consequences.
Read the full article here.