Trump Might Meet AI CEOs on September 29. Nobody Knows Who's Invited Yet

President Donald Trump, House Speaker Mike Johnson, and tech CEOs will reportedly meet on artificial intelligence (AI) on September 29, according to the latest reports. 

The meeting comes at a time when the AI safety debate has gained traction. Leading AI companies have warned of the technology’s risks, while other industry leaders question calls to slow down.

Anthropic’s Call Meets a Cybersecurity Skeptic

Earlier this month, Anthropic researcher Jacob Coxon resigned, alleging that Anthropic and OpenAI were not acting responsibly. His colleague Evan Hubinger put the risk of AI killing all humans within a decade at more than 10%.

Days later, Anthropic CEO Dario Amodei published an essay urging the industry to slow how fast AI capabilities improve. OpenAI’s Sam Altman and SpaceX’s Elon Musk backed the call.

However, Palo Alto Networks CEO Nikesh Arora questioned the motive. On X, he described the slowdown push as a stealthy commercial strategy. He later told CNBC he believed the labs wanted sympathy from regulators and a way around liability.

Arora now calls pacing unrealistic, arguing that some developers will keep building at the frontier.

“I think more than likely that some people will jump the gun, and you’ll have still people developing at the frontier, which means we shouldn’t try and stop the frontiers because they’re the most responsible people,” Arora said.

Arora also dismissed the 10% extinction estimate as highly unlikely.

Follow us on X to get the latest news as it happens

Are AI Labs Actually Pumping the Brakes?

Amodei’s essay said pacing would not mean halting training, and the labs that backed it have kept releasing models since.

Anthropic itself launched Claude Opus 5.5 on September 22, calling it its first model since urging a slower frontier. The company said outside evaluators, including METR, tested it before release.

OpenAI released GPT-6 Sol and Luna the same day. The pair followed the flagship GPT-6 Astra, which launched on September 4.

Google has also kept shipping. It released Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking on September 15.  Separately, DeepMind’s Koray Kavukcuoglu revealed that Gemini 4 has entered post-training.

As those models reach users, the Trump administration has weighed in on who should answer for their safety. On Monday, Treasury Secretary Scott Bessent said AI developers should take responsibility instead of expecting a federal liability shield. 

That stance leaves little room for the liability protection Arora suspected the labs wanted. No full guest list or detailed agenda for the September 29 meeting has surfaced so far. That leaves open whether talks will center on regulation, industry safeguards, US competitiveness, or some mix of them.

Subscribe to our YouTube channel to watch leaders and journalists provide expert insights