Logo via Wikimedia Commons; treatment-A cover, license to verify on approval
Altman acknowledged public fears about rapidly advancing AI but argued industry self-regulation is the path forward, while ruling out an OpenAI IPO for now
Sam Altman wants you to know that he’s scared of AI too. He just thinks you should let him handle it.
Speaking at Salesforce’s Dreamforce conference on September 15, the OpenAI CEO made a case that sounds like asking the fox to design the henhouse security system: the AI industry should be trusted to regulate itself on safety because its leaders understand the stakes better than anyone else.
“The world should trust that we are going to do the right thing because it’s the right thing and we feel the magnitude of this,” Altman told the audience. He added that he is “very confident in our industry’s ability to do this safely.”
The self-regulation sales pitch #
Altman’s remarks didn’t happen in a vacuum. Days before the Dreamforce appearance, Anthropic CEO Dario Amodei published an essay calling for a deliberate slowdown in advanced AI model development. Altman publicly endorsed the concept of “pacing the frontier” on X shortly after Amodei’s piece dropped.
Meta’s Mark Zuckerberg and Nvidia’s Jensen Huang also spoke at the event, reinforcing similar themes around industry-led safety standards.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
The backdrop here matters. OpenAI, Anthropic, and Google DeepMind have reportedly been engaged in safety standards dialogues since July 2026. These conversations have centered on independent evaluations, risk management frameworks, and shared benchmarks for what constitutes responsible development.
Altman went further than vague promises, stating that OpenAI is now preparing explicit safety cases ahead of its frontier training runs. He also drew a specific line in the sand, ruling out any scenario with a 10% or greater probability of catastrophic risk as flatly unacceptable.
Pacing, not pausing #
One of the more interesting tensions in Altman’s comments was his careful distinction between slowing down and stopping. “Pacing” development, in his framing, doesn’t mean pulling the emergency brake on AI progress. It means being more deliberate about how and when frontier models get deployed.
No IPO, more safety spending #
Perhaps the most commercially significant moment came when Altman indicated that an IPO for OpenAI would be “ill-advised” right now, given the company’s current focus on safety infrastructure.
The logic tracks in a cynical way. Going public would subject OpenAI to quarterly earnings pressure and shareholder demands for growth, exactly the kind of forces that make “pacing the frontier” harder to sustain. Staying private gives Altman more latitude to invest in safety research without explaining to Wall Street why those expenditures don’t immediately translate to revenue.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our