Jensen Huang predicts cybersecurity as AI's next big application

Jensen Huang, CEO of Nvidia, widely regarded as the leading company in AI chips, has named cybersecurity as the next major market for AI.

Speaking at a Goldman Sachs Group technology conference in San Francisco on Thursday local time, Huang predicted that "cybersecurity is likely to be the next major application area for AI," Yonhap reported, citing Bloomberg.

Huang said the advancement of AI models is accelerating the automation of computer programming, which in turn is speeding up both the exploitation of security vulnerabilities and the development of patches to fix them — a shift he said is transforming the cybersecurity industry as a whole.

"What better way to create demand than to create the problem?" Huang said, suggesting that the advancement of AI is making cybersecurity more vulnerable while simultaneously boosting demand for AI itself.

However, on the so-called "fear marketing" that plays up cybersecurity crises, he added, "Of course, there is a responsible way to do this, and there is a less attractive way."

The stance appears aimed at reassuring investors that spending on AI data centers will continue going forward.

Nvidia has also projected that demand for its AI chips will remain strong, as it seeks to ease concerns over talk of an AI bubble.

'Adopting AI at a company needs the same safeguards as hiring a new employee'

Huang also stressed AI-related safety earlier this month.

He said companies adopting AI should put in place thorough safeguards, just as they would when hiring a new employee.

Huang made the remarks in a video keynote address at "Equinix Horizon," an event hosted by digital infrastructure company Equinix on Sept. 2 local time at Midway in San Francisco, Yonhap reported.

"Introducing AI into a company's internal systems is like hiring a new employee," he said.

Just as a new employee should not be given immediate, unrestricted access to all of a company's critical files or internal tools upon joining, AI systems must also be granted clearly defined access privileges, he said.

He said the AI of the future will not remain confined to a single location but will shift toward a "highly distributed" structure spread across multiple data centers and cloud environments. Security controls are essential to prevent unauthorized data leaks from AI in such a fragmented network environment, he added.

Huang also said that in a future where distributed AI runs on fragmented data, infrastructure capable of securely tying these systems together will be essential.

yul@heraldcorp.com