Intel and Mirror Security Partner to Protect Autonomous AI Agents

Intel and Mirror Security today announced a collaboration to make autonomous AI agents more secure. This alliance brings together Intel’s advanced security features, including the Intel® Tiber™ Trust Authority and Confidential Computing, with Mirror Security’s AI behavior monitoring and encryption technologies. Together, they offer a powerful solution to help organizations securely deploy autonomous AI systems.

As AI agents are increasingly used to make decisions, take actions, and interact with sensitive systems, traditional security measures are no longer enough. These agents operate independently and often in unpredictable environments, making it difficult to control or verify their actions. The new solution from Intel and Mirror Security addresses this challenge by protecting AI systems not just when data is stored or shared, but also during processing — when decisions are being made.

“Our biggest concern with autonomous agents is ensuring they make the right decisions while operating freely,” said Pankaj Thapa, CEO of Mirror Security. “It’s like letting people move freely in a building while making sure they follow the rules. Our platform sets the rules, and Intel’s hardware confirms they’re being followed.”

Responsible Agentic AI

The solution offers several key features. It provides cryptographic proof that an AI agent’s actions stay within approved limits. It ensures the agent’s data and code are processed in secure, hardware-isolated environments using Intel SGX and Intel TDX. It also sets up guardrails to prevent unauthorized actions while allowing the agent to operate independently. Additionally, the system keeps a secure, unchangeable record of every decision the agent makes — useful for compliance and audits.

“As AI shifts from passive tools to active agents, we need stronger security for the moments when critical decisions are made,” said Purnam Sheth, VP and GM of Trust and Security Products at Intel Corporation. “This joint effort offers a way to build AI systems that organizations can trust — secured during every stage of decision-making.”

The solution can be used across multiple industries. In financial services, it helps build more secure autonomous trading agents with risk limits that can be verified. In healthcare, it protects patient data and supports trustworthy clinical trials. In enterprise settings, it supports document processing and complex workflows involving multiple AI agents. For critical infrastructure, it ensures monitoring systems follow safety protocols that can be verified.

The combined solution is available now for enterprise customers. It works on infrastructure enabled with Intel SGX, and support for Intel TDX will be available in June 2025. Mirror Security’s platform is available through annual subscriptions with enterprise support.