Bill Gates Tells Congress ‘Self-Regulation Isn’t Enough’ as He Warns AI Could Enable a Billion Deaths

Tech founders lobbying for government restraint on their own inventions is rare but not new; compare to J. Robert Oppenheimer’s 1945 efforts to curb nuclear proliferation and to Alfred Nobel’s 1895 will creating the Peace Prize after inventing dynamite. Each episode reflects a recurring pattern: an explosive breakthrough is commercialised faster than political institutions can react, prompting the inventors themselves to sound alarms. Gates’ reversal from the 1990s Microsoft antitrust battles—when the firm fought regulation tooth-and-nail—signals a structural shift: AI’s dual-use nature, like nuclear energy in the 1950s or genetic engineering after the 1975 Asilomar conference, pushes even profit-minded leaders to seek external rules. On a 100-year horizon, today’s debate could determine whether AI follows the trajectory of nuclear arms control (treaties, inspection regimes, partial successes) or social media (belated, piecemeal oversight). If policymakers seize the moment, this could be a pivot toward global tech governance; if they stall, the warning may echo earlier missed chances—such as the lack of internet privacy laws in the 1990s—that proved impossible to retro-fit later.