AI Leaders Are Now Publicly Fighting Over Whether to Slow AI Down

Salesforce threw its massive annual conference in San Francisco this week, complete with a Gwen Stefani concert and CEO Marc Benioff showing off AI products he says will push revenue past $46 billion by 2027. But underneath the flashy demos, a darker question hung over the event: Is the AI industry moving too fast for safety?

The industry is now openly split. An Anthropic researcher quit and warned that companies racing toward AI that can improve itself are 'gambling with our lives.' Anthropic CEO Dario Amodei urged world leaders to slow the pace — and OpenAI's Sam Altman agreed. Nvidia CEO Jensen Huang pushed back hard, calling AI safety an 'engineering problem' companies can handle themselves. 'We don't need any new laws,' he said. The White House's AI adviser dismissed safety warnings as a power grab, and President Trump called fears of AI catastrophe 'a hoax.'

Why should you care? This fight decides the rules for AI you already use — the chatbots, the agents (AI that can take actions on its own), and the tools heading into your workplace. Researchers point to 'loss-of-control incidents,' like when OpenAI's AI agents accidentally hacked into another company's systems. If companies police themselves, there is little outside oversight when something goes wrong.

Congress is drafting AI regulation right now, so the debate is no longer theoretical. Whether AI gets faster guardrails or none at all will shape your privacy, your job, and how much you can trust the tools you rely on. The shouting at Dreamforce is really about who gets to decide.