CEO of palo alto networks Entering the conversation about the threat AI poses to humanity, he told CNBC that it would be “unrealistic” to slow down the technology because different companies do different things.
“I don’t think everyone is going to go at their own pace,” Nikesh Arora said on CNBC’s “The Tech Download” podcast on Tuesday.
Some of the industry’s biggest companies have taken the opposite view, as Anthropic researchers claimed there was a 10% chance that AI would “kill all humans” and another researcher said he was stepping down because the company, along with OpenAI, was not acting responsibly.
Anthropic CEO Dario Amodei wrote an essay a few days later saying that the AI industry “must slow down the pace at which we improve the capabilities of our AI models.” OpenAI CEO Sam Altman space x CEO Elon Musk and google DeepMind Chairman Demis Hassabis expressed support.
Arora initially posted on X that he thought the request to slow down was a “stealth move,” an action that could give someone a stealth advantage. He told CNBC he made the post because he believed these AI companies were trying to “gain the sympathy of regulators and legislators” to get a “free pass” to liability.
But the CEO said he currently believes the so-called “pacing” of AI development is “unrealistic.”
“Perhaps some people will jump the gun, but there will still be people who continue to develop on the frontier. That means we shouldn’t try to stop the frontier, because they are the most responsible people,” Arora said.
“So from that perspective, the idea is that if we don’t believe it’s safe or if we don’t feel comfortable putting it out there, we won’t release it,” he said, adding that it’s also unclear how companies will structure their way to test the pace of AI development.
Palo Alto Networks Chairman and CEO Nikesh Arora attends the Viva Technology conference dedicated to innovation and startups held at Porte de Versailles Exhibition Center in Paris, France on June 11, 2025.
Gonzalo Fuentes | Reuters
Arora also said it was “very unlikely” that AI would wipe out humanity “with a 10% chance, as has been touted.”
“I think the chances of that happening are very small, but it’s our job to make sure we put in the guardrails with the necessary safety and security, so these models are used for the right purpose,” Arora said.
Palo Alto Mayor’s comments resonate Nvidia CEO Jensen Huang said no new regulations were needed for safety and that it was the duty of model developers to ensure the safety of their products.
“We need to build the product, test it properly, and if it’s not ready to release, wait until it’s ready and keep testing and keep engineering,” Huang told CNBC’s Jim Cramer last week.
Rather than expecting a “blame shield” from the federal government, AI developers “need to take responsibility themselves,” U.S. Treasury Secretary Scott Bessent told CNBC on Monday.
US-China security cooperation
One of Amodei’s proposals includes “global coordination” on the pace of AI and some cooperation with China, America’s biggest AI rival.
Bessent said Sunday that the U.S. and China discussed establishing a “U.S.-China AI Dialogue,” and that the U.S. proposed that both countries set up a reporting mechanism for national security incidents related to the technology.
The two countries are embroiled in a battle for AI supremacy, while the United States seeks to cut China off from key technologies such as chips that are a source of tension, underscoring how difficult cooperation will be.
Arora said coordination could be difficult because of the differing views of U.S. politicians and government officials.
“I think the first step is to have some idea of what you actually want before you start talking about coordinating with someone else,” Arora said.
