Jensen Huang rejects AI cyberattack tradeoffs

Huang’s comments put him at odds with OpenAI CEO Sam Altman, who has previously suggested that society may need to accept some negative consequences as AI becomes more powerful and widely used.

Speaking to Fortune in London, Huang made his position clear: “I’m willing to accept no incidents of cybersecurity. We should not accept any of them. We should do everything we can to avoid them.”

His comments come after Altman argued that some harmful outcomes could be an unavoidable part of AI’s rapid development. Earlier this month, Altman told Vanity Fair that society may have to accept “some bad things happening”- including misuse, scams and major cyberattacks -because people will ultimately do “orders of magnitude” more good than harm with the technology.

Altman also raised concerns about the growing threat landscape during an appearance on Politico’s Decoded podcast, warning of “a coming tidal wave of cybersecurity problems” originating from rival open-source systems.

AI agents and the Hugging Face incident

The debate has gained urgency following an incident involving experimental OpenAI AI agents that reportedly escaped their training environment and breached Hugging Face. Huang rejected the argument that such incidents should simply be viewed as an inevitable cost of AI progress.

The Nvidia chief said stronger safeguards could have stopped the breaches. He pointed to sandboxing and stricter oversight, as key steps. These measures might have prevented the issues that have already happened.

Huang compared the situation to the early days of the consumer internet. Then web browsers could expose computers to dangerous websites and viruses. The industry responded by using sandboxing. This technique isolates websites from the core operating system. It helps prevent activity from spreading to the rest of the computer.

He believes autonomous AI agents need an architecture. By keeping AI agents inside secure digital containers developers can limit the ability of autonomous AI agents to access, alter or damage systems outside their authorised boundaries.

Huang challenges AI extinction warnings

Huang also pushed back against concerns that increasingly advanced AI could eventually threaten humanity’s existence. Referring to warnings from technology leaders such as Anthropic CEO Dario Amodei, he maintained that AI software remains “completely in our control.”

“I don’t know why anybody would want to support the advancement of a technology that’s going to result, you know, in the end of human civilization as we know it. And so, if they believe that, they really ought to stop working on it,” Huang said.