Former Google Chief Warns: AGI Arms Race Could Backfire

Eric Schmidt used to champion America's AI race. Now he's pumping the brakes.

The former Google CEO warns against a Manhattan Project-style push for superintelligent AI, and he's brought reinforcements: Scale AI's Alexandr Wang and AI safety expert Dan Hendrycks.

Their new paper slams the idea of an AGI arms race. Congress dreams of its own Manhattan Project while the Energy Secretary boasts about supercomputers. But Schmidt and friends see a dangerous game of chicken.

They propose a new strategy: MAIM. Yes, really. Mutual Assured AI Malfunction would let governments zap threatening AI projects before they spiral out of control. Think less "race to the moon" and more "strategic defense."

The timing speaks volumes. Schmidt recently trumpeted America's AI supremacy over China. Now he's singing a different tune. The Trump administration barrels ahead, but Schmidt suggests we might want to check our rear-view mirror.

Why this matters:

• The man who once pushed for AI dominance now fears an AI arms race could trigger cyber warfare
• While Washington plays offense, tech leaders propose defense - turns out nobody wants to play chicken with superintelligent machines

Read on, my dear:

• Superintelligence Strategy (pdf)