Bill Gates shifts stance: once AI optimist now warns of "worst source of injustice"

After months of relative silence, Bill Gates has published a lengthy manifesto signaling a dramatic reversal in his views on artificial intelligence. Where the billionaire once championed AI's potential, he now voices alarm about its trajectory—and is actively pushing back into the debate to reshape how the world manages the technology's risks.

A stark reversal#

Gates' latest essay, titled "The turbulent AI era is here. The choices we make now are critical," abandons the hopeful framing of his 2023 piece on AI. He characterizes the coming transformation as potentially "the worst source of injustice" humanity has created, or conversely, "the greatest equalizer ever invented"—the outcome, he argues, hinges on decisions made today.

The roughly 6,000-word piece, released after Gates stepped back from public AI discussions, reflects deep concern that global institutions are fundamentally unprepared for what lies ahead. "We are not preparing for it," he writes, warning that policymakers underestimate both the pace and scale of AI's societal impact.

The labor crisis ahead#

Gates devotes considerable attention to employment disruption, forecasting that artificial intelligence will permanently eliminate vast swaths of jobs across both white-collar and blue-collar sectors. Facing potential mass unemployment and social destabilization, he urges political leaders to act preemptively rather than reactively.

His proposed remedies are unusually specific for a tech figure: tax AI tokens and robots to discourage companies from replacing workers, redirect those revenues into social safety nets, and formally designate certain roles as "Human Reserved"—positions protected for people. Gates emphasizes the urgency, noting that waiting until "unemployment rises sharply, communities are hurting, and public trust has eroded" will be too late.

Building institutions for a cross-cutting threat#

Gates argues that existing governance structures are ill-equipped to handle AI's complexity. Unlike technologies that fit neatly into single policy domains, artificial intelligence simultaneously touches tax policy, labor, health, national security, and education—yet no institution sees the whole picture.

He calls for two institutional shifts: domestic bodies within governments that coordinate AI strategy across agencies, and a novel international organization to manage transnational risks. While Gates draws potential inspiration from frameworks governing nuclear power, aviation, and ozone protection, he offers few specifics on what such an organization would resemble or how countries—particularly the US and China—would cooperate within it. He does stress the need for speed, urging nations to begin building these structures immediately.