Leaders of the world’s most advanced artificial intelligence companies are helping create an “unnaturally negative” narrative about the technology, Palo Alto Networks CEO Nikesh Arora said at Semafor’s Silicon Valley & The World event.
In an interview with MIT Technology Review’s Mat Honan, Arora argued that attention to risks is overshadowing the potential benefits of the fast-moving technology. He said it was strange that leaders working on frontier large language models sometimes focus on AI’s dark side, when they should also be advocating its benefits.
Arora described AI as a transformational technology that could improve many people’s lives and make knowledge more widely accessible. As an example, he said a villager in India or elsewhere could gain access not only to multiple websites, but also to summaries of their contents and the best available information.
Data centers and efficiency
The CEO also challenged concerns that growing domestic resistance to data center construction could slow innovation. He argued that limits on computing capacity would instead encourage companies to develop more efficient approaches.
Scarcity breeds innovation, Arora said: if only a limited number of data centers can be built, more companies could offer compression technologies that make token processing more efficient.
Growing cybersecurity risks
Arora nevertheless acknowledged that AI’s expanding capabilities will inevitably intensify cybersecurity risks by making software vulnerabilities easier to discover. Weaknesses in flawed software written over the past 50 years, he said, are now more likely to come to light.
He warned that the more vulnerabilities are discovered and exploited, the greater the risk that systems will be forced offline as companies rush to patch longstanding flaws. He added that he saw neither a panacea nor a quick solution, and that the problems would require sustained work to overcome.



