Palo Alto CEO: Slowing AI Development 'Impractical,' AI Firms Seek Regulatory 'Sympathy'

According to a report by Benzinga, Nikesh Arora, CEO of Palo Alto Networks (PANW-US), has commented on recent discussions about slowing down AI development, suggesting it may be a strategy used by some companies to win 'sympathy' from regulators.

Arora appeared on CNBC's podcast program 'The Tech Download' on Thursday, addressing the ongoing debate over whether AI could pose a threat to humanity. He stated that AI companies are trying to 'gain sympathy from regulators' and seeking a 'free pass' when it comes to responsibility. He also described the idea of controlling the pace of AI development as 'impractical'.

The Palo Alto Networks CEO advised that companies should not release products unless they are confident in their safety. He also expressed skepticism about whether businesses can establish a system to monitor and control the speed of AI advancement.

Arora dismissed claims that AI could lead to human extinction, calling such scenarios 'extremely unlikely' and the probability 'minuscule.' He emphasized that safety and protective measures must be integrated into AI models to ensure responsible use of AI.

Last week, Arora responded on X to Anthropic CEO Dario Amodei's call for the entire industry to slow down AI development, labeling it a 'ninja move.' He argued that AI safety measures ultimately serve as a business strategy: companies can demonstrate 'duty of care' while encouraging competitors to adopt the same standards, thereby establishing a first line of defense across the industry.

The speed of AI development has recently become a hot topic. Earlier this month, Shyam Sankar, Chief Technology Officer at Palantir Technologies (PLTR-US), voiced concerns about the politicization of AI safety, accusing a group known as 'Effective Altruists' of attempting a 'coup' by deciding the pace of technological progress for everyone.

Shortly afterward, Jensen Huang, CEO of NVIDIA (NVDA-US), opposed new antitrust rules that would allow AI companies to coordinate in slowing the development of more powerful AI models. He acknowledged that AI safety is a 'real issue' but maintained that companies should not release unsafe products.

Notably, Palo Alto Networks earlier this week launched a new Unit 42 service that leverages AI models from Anthropic and OpenAI to continuously identify security vulnerabilities, test attack paths, and prioritize remediation.

The service was tested over six months in collaboration with more than 100 customers and backed by $17 million in funding. The results showed that it identified a year's worth of accumulated security exposure issues within enterprises in just three weeks. Security exposures were found in all assessed customers.