Nvidia is rolling out new NIM microservices, or small standalone services, to help enterprises bring additional control and safety measures to their AI agents.
One of these new NIM features targets content safety and works to prevent an AI agent from producing harmful or biased results. A third new service helps the AI agent protect against jailbreak attempts or remove software restrictions, but otherwise keeps conversations focused on approved content.
This is part of three new NIM microservices. Nvidia NeMo GuardrailsNvidia's collection of existing software tools and microservices to help companies improve their AI applications.
“By using multiple lightweight and specialized models as shields, developers can cover the gaps that can occur when only generic global policies and protections exist — a one-size-fits-all approach cannot securely control complex agent-driven AI workflows,” he says. Press release He said.
AI companies are likely to start realizing that getting businesses to adopt their AI agent technology won't be as simple as initially thought. People like Salesforce CEO Marc Benioff recently. Estimated. Over a billion agents will flee Salesforce alone in the next 12 months; The reality will be slightly different.
A recent study by Deloitte Estimated. About 25% of enterprises are already using AI agents or expect to be by 2025. It is also predicted that by 2027, half of businesses will use agents. Businesses are clearly interested in AI agents, but it shows that AI technology is not being adopted in the same way as innovation is happening in the AI field.
Nvidia hopes initiatives like this will make using AI agents seem safer and less experimental. Time will tell if that is true.