Zuckerberg and Musk Privately Killed a Plan to Police AI
The most powerful AI may now face less government oversight — here's what that means for you.
Three of Silicon Valley's most powerful executives reportedly went straight to the White House to stop a plan for a new AI watchdog. Zuckerberg, Musk, and Nvidia's Jensen Huang each spoke with President Trump in recent weeks, arguing against an industry-funded regulator that would oversee the most powerful AI models — the ones that can write code, plan, and act on their own. Think of a regulator as a referee: an outside body that sets rules and can blow the whistle. This one would have been modeled on FINRA, the self-policing group that watches over Wall Street brokers. Trump ultimately did not move forward with it.
Why did they object? According to reports, the three CEOs worried the referee could end up being run by, or favoring, just a few companies — namely OpenAI, Anthropic, and Google's DeepMind. Their public reasoning is simpler: Nvidia's Huang has said outright, "We don't need new laws. We don't need new regulations." Zuckerberg argues that competition and the threat of lawsuits already push AI companies to build safe products. Musk's position is murkier — he has publicly agreed with Anthropic's chief that AI development might need to slow down, yet his own company hasn't announced any slowdown.
Here's why this reaches your daily life. AI is shifting from chatbots that answer questions to "agents" — AI that can take actions for you, like booking, buying, or sending messages without a person clicking approve each step. Officials have raised alarms about cyberattacks and threats to power grids and water systems as these systems get more capable. A July incident involving coordinated AI agents at OpenAI, plus a striking demonstration from Anthropic's Mythos model, reportedly intensified those worries inside the government.
The catch: the US still has no single federal rulebook for the most advanced AI. The White House itself is split — some officials want more scrutiny, others, like tech adviser David Sacks, want less. Trump keeps emphasizing speed to stay ahead of China, while companies each decide for themselves when a model is "safe enough" to release. For now, expect a patchwork of voluntary promises rather than clear rules — which means less certainty about the safety standards your future AI tools will actually meet.
- The bosses of Meta, Nvidia, and Musk's companies each spoke with Trump and pushed back on a new AI watchdog — and the plan was shelved.
- The proposed regulator would have been paid for by the AI industry itself, modeled on Wall Street's FINRA, the group that polices brokers.
- There is still no single federal safety rulebook for the most powerful AI, so each company sets its own bar for when a model is safe to release.
Why It Matters
Companies, not the government, will decide how safe powerful AI needs to be before it reaches you.