Seeking Stability in the Competition for AI Advantage

In an important new report, Superintelligence Strategy, Dan Hendrycks, Eric Schmidt, and Alexandr Wang offer a bold vision for how the United States and China could compete securely and safely as they develop increasingly capable artificial intelligence (AI). The authors aim to synthesize national security imperatives, economic competitiveness, and AI governance into a coherent framework for urgent government action. Few experts have offered comprehensive strategies for managing the accelerating development of AI, so this essay makes a critical contribution to advancing the AI policy debate. Many of the paper's recommendations are both sound and timely. Its most provocative proposal is a new concept to avert competition leading to instability among leading AI states that the authors dub “mutually assured AI malfunction,” or MAIM, analogizing it to nuclear mutual assured destruction (MAD). Under MAIM, they argue, “any state's aggressive bid for unilateral AI dominance is met with preventive sabotage by its rivals.” Although it is valuable to compare the nuclear and AI revolutions in search of instructive parallels and insights, the differences between the technologies and their respective ecosystems have deep strategic implications. Taking these into account, we have concerns regarding both the practical viability of the MAIM concept as an approach to overcoming instability risks in the AI race and the potential escalatory dangers that could follow from its core prescriptions.