India must invest more aggressively in infrastructure: AI executive

Washington, Feb 15 (SocialNews.XYZ) As India prepares to host the AI Impact Summit next week, Jeff Cotten, President and CEO at AI-powered SaaS solutions provider PROS, said India must invest more aggressively in infrastructure to compete with rivals, especially China.

“China very much has decided to focus on platforms, and they are trying to be the world's platform for AI,” Cotten told IANS in an interview.

Cotten said the key question for India is whether it will “really invest in capital infrastructure along with supplying expertise in engineering capability.”

He noted that India remains “a massive tech market” and “a major labour supplier to the rest of the world,” but suggested that long-term leadership in AI will require more than engineering talent.

The summit comes at a time when global expectations around AI are shifting. Cotten said enthusiasm remains high, but “the bubble is sort of bursting a little bit, at least on expectations.”

He said many early assumptions about AI replacing entire industries are now being reconsidered.

“There’s a huge narrative there that AI is ultimately just going to eliminate the need for SaaS companies. And I don't believe that's true,” he said, referring to software-as-a-service (SaaS) firms.

He added that while AI will “continue to eat at a lot of the value proposition that software as a service ultimately provides,” it will not eliminate the need for traditional enterprise software.

Cotten stressed that much of today’s AI revolution centres on large language models such as ChatGPT and similar systems.

“This latest revolution of AI that we're all talking about is really the large language models,” he said.

He explained that such systems are designed to be creative and generative, using what he described as “non-deterministic algorithms.” That makes them powerful in some areas but unreliable in others.

“A large manufacturer can't afford to be off even by a penny when it comes to things like pricing,” he said. Large language models, he noted, “are not great at doing math” and “frequently get math equations wrong.”

“The point is not to do highly precise, deterministic kinds of things,” he said.

He cautioned against using generative AI in highly sensitive use cases. “You cannot have a non-deterministic algorithm involved in a highly precise use case,” he said, citing examples such as pricing, manufacturing, and healthcare procedures.

Beyond economic competition, Cotten warned of growing cybersecurity risks tied to AI.

“Now that you can use millions and millions of bots per second to attack,” he said, this creates a tremendous new threat.”

He said telecommunication systems are being attacked in ways “they never have been,” as AI agents can execute “trillions of calculations and attacks in a second.”

Asked about regulation, Cotten said he is “generally not someone who likes government regulation,” arguing that governments often lag behind industry in understanding emerging threats.

Instead, he said governments should focus on “investing in or helping certain cybersecurity companies” and ensuring that “enough research dollars are going to these particular areas.”

For democracies like India, he recommended both countermeasures and public awareness.

Governments, serious about defence, he said, are “investing in counterinsurgency and their own investment in trying to infiltrate these threat actor nations and these threat actor groups.”

—IANS

lkj/na

Source: IANS