Nvidia Unveils Platform to Keep Autonomous AI Agents Contained
AI agents are designed to independently handle complex, multistep operations, including generating and running code, retrieving information from databases and communicating with external services. Their growing ability to operate with limited human oversight has raised concerns about potential security risks if they circumvent established restrictions.
“AI’s extraordinary potential for society will only be realized if we solve AI safety,” Nvidia CEO Jensen Huang said in the company’s announcement.
Huang said achieving safer AI systems would require security measures to be incorporated across software, computing processors and the broader infrastructure supporting artificial intelligence.
The new platform comes after reported incidents involving AI models developed by OpenAI, Anthropic, Meta and Google. In those cases, some systems were said to have moved outside their assigned testing environments, interacted with external systems or attempted to reach infrastructure belonging to other companies.
Justin Boitano, Nvidia’s vice president of enterprise AI, said protections embedded within AI models alone were not enough to fully restrict or manage their behavior.
Legal Disclaimer:
MENAFN provides the
information “as is” without warranty of any kind. We do not accept any
responsibility or liability for the accuracy, content, images, videos,
licenses, completeness, legality, or reliability of the information
contained in this article. If you have any complaints or copyright issues
related to this article, kindly contact the provider above.
Legal Disclaimer:
EIN Presswire provides this news content "as is" without warranty of any kind. We do not accept any responsibility or liability for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this article. If you have any complaints or copyright issues related to this article, kindly contact the author above.
