[DigitalToday reporter Jinju Hong] Palo Alto Networks CEO Nikesh Arora (니케시 아로라) said calls to slow the pace of artificial intelligence development are unrealistic.

On Sept. 24 local time, CNBC reported that Arora said an industry-wide effort to rein in speed would be hard to make work because each company makes different judgments and takes different actions.

In a CNBC podcast interview, Arora said not everyone would match their pace on their own, and he saw it as likely that some companies would move first. For that reason, he argued companies developing AI on the front lines should not be made to stop across the board. Instead, he said products should be released only when companies can be confident they are safe before launch.

The remarks came as industry debate has intensified over AI safety and the pace of development. An Anthropic researcher previously made a claim to the effect that there is a 10 percent chance AI could kill all humans, and another researcher said the person left the company, saying Anthropic and OpenAI were not acting responsibly. Anthropic CEO Dario Amodei later published a post saying the pace of AI model performance improvements should be slowed. Sam Altman, Elon Musk and Demis Hassabis also expressed support.

Arora said such calls for slowing down could be seen as an attempt to win sympathy from regulators and lawmakers. He previously wrote on X, formerly Twitter, calling the argument a "ninja move". In the interview, he explained he wrote that because it looked as if AI companies were trying to win sympathy from regulators and politicians and effectively receive immunity on questions of responsibility.

He stressed that safe deployment standards matter more than the slowdown itself. "If you judge it is not safe, you should not launch the product," Arora said. He added it was also unclear how to create standards or testing systems that could verify the pace of AI development.

He assessed the possibility as low when it comes to warnings that AI could lead to human extinction. Arora said such an outlook was "very unlikely" and he did not agree with the claim of a 10 percent chance of eliminating humanity. "That probability is small, extremely small," he said, adding it is the industry's role to build guardrails and safety and security measures so models are used for the right purposes.

The view overlaps with recent comments from other industry figures. Nvidia CEO Jensen Huang said model developers should be responsible for the safety of their products rather than introducing new regulation. U.S. Treasury Secretary Scott Bessent also said AI developers should take responsibility themselves rather than expect protections from the government.

Among the measures proposed by Dario Amodei was global coordination over the pace of AI development and a certain level of cooperation with China. Arora said the United States should first sort out what it wants domestically before such discussions. "There are differences in views even between U.S. political circles and corporate executives," he said. "Before talking about coordination with other countries, we should first put some order into what we actually want."

The United States and China are competing for leadership in AI. The United States has sought to block China's access to core technologies such as semiconductors, and the issue has been a backdrop to tensions between the two countries. In this situation, whether international coordination over the pace of AI development will work is expected to remain an important issue.