AI Bosses Keep Begging for Rules — But Nothing Ever Happens

In the past few days, some of the richest people in tech all said the same thing in public: AI is getting dangerous, and governments should step in and slow it down. That group includes OpenAI's Sam Altman, Anthropic's Dario Amodei, Google DeepMind's Demis Hassabis, Microsoft's Satya Nadella, and Elon Musk. When the people who make money from something announce it needs to be controlled, it's reasonable to raise an eyebrow.

And it turns out they've been saying this for years. Elon Musk told a room full of US governors in July 2017 that AI needed regulation immediately — at a time when he had already put $38 million into OpenAI and held stakes in other AI companies. He'd spent three years calling AI "more dangerous than nukes" and "summoning the demon." In 2018, Mark Zuckerberg told lawmakers he was "actually not sure we shouldn't be regulated," in the middle of the Cambridge Analytica privacy scandal. The pattern is old: warnings first, rules later (or never).

The worry goes back further than any of them. Author Samuel Butler wrote in 1863 about machines replacing humans, and computing pioneer Alan Turing warned in 1951 that intelligent machines could "outstrip our feeble powers." In 2000, Sun Microsystems cofounder Bill Joy compared self-replicating robots to nuclear weapons. In 2009, Microsoft researcher Eric Horvitz gathered top AI scientists to talk about constraining autonomous systems.

So what does this mean for you? Mostly, watch what they do, not what they say. A public call for safety rules is cheap — it builds trust, shapes whatever laws eventually get written, and costs nothing today. Nearly a decade after Musk's first big warning, there are still few meaningful AI laws anywhere. Until real rules arrive with real penalties, treat the alarm-sounding as a signal about politics, not a guarantee that anything is about to change.