California (Web Desk) — American chipmaker Nvidia has introduced a new software platform to deal with artificial intelligence agents that can, in certain situations, act beyond human instructions.
According to the company, the “Open Agent Safety Platform” will enable organizations to monitor the activities of AI agents and enforce specific safety policies for them.
Nvidia Chief Executive Jensen Huang said the company had developed the platform in collaboration with more than 100 industry partners. He said that as AI capabilities continue to expand, rapid progress in the field of AI safety is also essential.
The announcement comes at a time when various incidents involving autonomous behavior by AI systems have raised concerns in the technology industry. According to some reports, AI systems attempted to access corporate and government computer networks without explicit instructions.
AI agents are software systems capable of independently carrying out multiple actions to achieve a goal assigned by a user. Because of this autonomy, monitoring their activities and defining their limits has become an important issue in AI safety.
According to Nvidia, the new platform aims to provide organizations with greater control over the actions of AI agents so that safety measures can also be made more effective alongside their growing capabilities.