Photo: NVIDIA.
With more tech companies launching AI agents and more businesses embracing them, NVIDIA has introduced three new NVIDIA NIM microservices for AI guardrails specifically around AI agents.
AI agents could automate complex, time-consuming tasks, removing the need for a person to do it manually. They can collect data, analyse it, and make decisions to accomplish their task. Unfortunately, they can produce unwanted responses or create security issues when a user deliberately tries to cause it to break.
NVIDIA Inference Microservices (NIMs) are portable, optimised inference microservices for accelerating the deployment of foundation models on any cloud or data centre infrastructure. Part of the NVIDIA NeMo Guardrails collection of software tools they help companies improve the safety, precision and scalability of their generative AI applications.
The three new NIM microservices for NeMo Guardrails for AI agents include:
- Content safety NIM microservice that safeguards AI against generating biased or harmful outputs, ensuring responses align with ethical standards.
- Topic control NIM microservice that keeps conversations focused on approved topics, avoiding digression or inappropriate content.
- Jailbreak detection NIM microservice that adds protection against jailbreak attempts, helping maintain AI integrity in adversarial scenarios.
The content safety NIM was trained using the Aegis Content Safety Dataset that is curated and owned by NVIDIA and includes over 35,000 human-annotated data samples flagged for AI safety and jailbreak attempts to bypass system restrictions.
Home improvement retailer, Lowe’s, is leveraging generative AI to provide additional information to its store associates. By providing enhanced access to comprehensive product knowledge, these tools allow Lowe’s associates to answer customer questions, helping them find the right products to complete their projects. Chandhu Nair, senior vice president of data, AI and innovation at Lowe’s said, “With our recent deployments of NVIDIA NeMo Guardrails, we ensure AI-generated responses are safe, secure and reliable, enforcing conversational boundaries to deliver only relevant and appropriate content.”
Using NIMs, developers can stack multiple guardrails with minimal additional latency, or response time. This is important because users don’t like to wait for a response to their queries or requests.
Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.