Multicloud security firm Fortanix Inc. today announced the public preview of Fortanix Armet AI, a new service that allows enterprises to create secure and compliant custom generative artificial intelligence applications to mitigate data exposure risk.

Armet AI has been designed to assist organizations that are seeking to take advantage of generative AI technology but don’t view off-the-shelf solutions as a viable option given the risk of security failures that expose sensitive data and intellectual property, among other concerns. Fortanix argues that alternatives, such as do-it-yourself solutions, can also be complex, costly and resource-intensive.

The new Fortanix offering addresses the challenges of deploying generative AI with a turnkey generative AI platform. The company says Armet AI ensures generative AI data pipelines have the highest levels of security and meet comprehensive regulatory and compliance standards, allowing organizations and users to gain instant, relevant answers using their organization’s own internal and sensitive data.

Armet AI is powered by Intel Confidential Computing, including Intel SGX, Intel TDX and Intel Tiber Trust Authority. The foundation protects data in use by performing computation in a hardware-based, attested, trusted execution environment, preventing unauthorized access or the ability to alter data.

“Generative AI represents a significant leap forward with the immense promise to reshape how organizations harness and act upon their data,” said co-founder and Chief Executive Anand Kashyap. “While it’s natural to be wary of emerging cyberthreats, regulatory pressures, and reputational risks, Fortanix Armet AI offers a fast path forward: secure, instant access to internal knowledge that propels informed decision-making without compromising compliance or trust.”

Key features of Armet AI, along with confidential computing, include role-based access control that helps organizations maintain strong data governance by enforcing fine-grained policies that dictate who can train, access and view specific datasets. The controls support regulatory compliance while minimizing the risk of unauthorized data exposure.

The new service features built-in AI guardrails that protect data and models by automatically redacting sensitive information and filtering harmful inputs or outputs. Doing so helps ensure safe and secure generative AI interactions, even in highly regulated or risk-sensitive environments.

Armet AI also includes REST APIs and prebuilt connectors to streamline data integration and collaboration across systems. Additionally, a secure chat interface and a generative AI security dashboard provide users with private access to internal knowledge and real-time monitoring to support proactive AI governance.

Image: SiliconANGLE/Reve