Zuckerberg, Musk, and Huang Teamed Up to Kill Their Own Industry's Safety Referee
Three of tech's most powerful — and usually competing — CEOs just agreed on one thing: they didn't want an AI regulator, even one their own industry would have funded and run.
Mark Zuckerberg, Elon Musk, and Nvidia's Jensen Huang separately called President Trump in recent weeks to reject a proposed oversight body for frontier AI models, modeled on ideas floated by Google DeepMind's Demis Hassabis.
Here's what was actually on the table, and why it collapsed:
→ The plan resembled a FINRA-style body — a standards-and-supervision organization funded by the AI industry itself, not a government agency
→ Zuckerberg, Musk, and Huang argued it would concentrate power in the hands of OpenAI, Anthropic, and Google — the very labs that would likely staff and fund it
→ The pushback landed the same week Anthropic's Dario Amodei called for a slower pace of AI capability gains, with Sam Altman and Hassabis publicly siding with more caution
→ Huang took a different tack at Salesforce's Dreamforce conference — arguing existing product-liability and safety laws are already enough: if a company isn't confident in a system, it simply shouldn't ship it
→ Musk proposed an alternative: rival labs auditing each other's models instead of a shared referee
→ White House officials told other AI executives that building internal consensus is difficult because opponents of tighter rules "call Trump directly" — meaning access, not argument, is deciding outcomes
→ Trump has publicly dismissed existential AI risk warnings as a "hoax" and argues heavy restrictions would hand China an advantage
Here's what stands out to me: even the mildest version of self-regulation — an industry group policing itself, with no government teeth — couldn't survive contact with the three companies that would have had the most resources to shape it in their favor. That's worth sitting with. A FINRA-style body isn't external oversight; it's still industry insiders setting the rules. If even that was too much for Zuckerberg, Musk, and Huang, it suggests the disagreement isn't really about how AI gets regulated — it's about whether it gets regulated by anyone with real authority at all.
The practical result: no shared AI rulebook. Just personal access to the White House deciding whose safety concerns get heard, and whose don't.
Do you think an industry-funded safety body would have actually worked — or was it always going to end up protecting the biggest players regardless of who built it?
#AI #TechPolicy #AIRegulation #Nvidia #Meta #ArtificialIntelligence #TechNews
— 𝔖𝔞𝔫𝔡𝔢𝔢𝔭 ℜ𝔞𝔦𝔷𝔞