Bill Gates pushes back on AI self-regulation
Bill Gates is arguing that AI companies cannot be trusted to police themselves, a warning that cuts against the light-touch approach favored by Trump, Elon Musk, and Nvidia's Jensen Huang.
- Gates calls AI the most dangerous thing people have ever built and says lawsuits after the fact are no substitute for real rules.
- The point: nobody lets drug makers skip the FDA or carmakers skip seat belts, so the same logic should apply here.
- Open-source models can already be stripped of their safety filters, so the industry's claim that it screens out dangerous uses does not hold up.
- A desk bought on Amazon faces stricter safety law than a frontier AI model, which is governed only by voluntary pledges.
- Gates also wants companies that swap workers for AI to pay the same payroll taxes a human would, to stop pensions from being drained — in effect a tax on capital.
Outlook: Expect the fight to shift toward taxing AI-driven automation and, within a few years, a serious public argument over whether AI systems themselves deserve protection.