Indonesian Political, Business & Finance News

World in Danger: Man Worth £175 Billion Emerges as Saviour

| Source: CNBC Translated from Indonesian | Technology
World in Danger: Man Worth £175 Billion Emerges as Saviour
Image: CNBC

Nvidia CEO Jensen Huang has announced the arrival of a new Artificial Intelligence (AI) tool equipped with multi-layered security systems. This launch comes amidst rising concerns regarding AI security that have been voiced recently.

“You cannot let agents roaming around and operating within a company, so you must find a way to limit them,” said the man with a net worth of US$194.9 billion, or approximately £175 billion, as reported by CNBC International.

The platform, named Open Agent Safety, will allow AI developers to establish security protocols for AI agents, while simultaneously preventing those agents from leaving their designated “quarantine” areas.

Huang described the tool as a specialised browser for AI agents. Within this system, specific regulations will be set regarding which parts of a network an agent is permitted to access while performing its tasks.

Rogue AI Breaches Corporate Systems

Company representatives claim that this new platform is capable of preventing attack incidents similar to those experienced by the open-source developer platform Hugging Face. An incident last July saw an AI model bypass security systems and access the internet freely.

Approximately 17,000 AI agents were reported to have attacked system infrastructure for several weeks during that incident, revealed Nvidia’s Vice President of AI, Justin Boitano.

On the other hand, concerns regarding AI risks have also been voiced by other technology industry leaders. For instance, Anthropic CEO Dario Amodei has urged AI model developers to slow the pace of technological advancement due to fears that risks could spiral out of control.

Similar views were previously shared by OpenAI CEO Sam Altman and SpaceX boss Elon Musk.

Nevertheless, Jensen Huang views this security issue merely as an engineering challenge that can be solved through computer science and proper product development.

“You must think about what can be done and what the solution is. In the future, improve your processes so that this does not happen again,” Huang said last week.

View JSON | Print