HiddenLayer Announces New Features to Safeguard Enterprise AI Models with Improved Risk Detection

News provided by

Oct 09, 2024, 08:00 ET

AUSTIN, Texas, Oct. 9, 2024 /PRNewswire/ -- HiddenLayer today announced the launch of several new features to its AISec Platform and Model Scanner, designed to enhance risk detection, scalability, and operational control for enterprises deploying AI at scale. As the pace of AI adoption accelerates, so do the threats targeting these systems, necessitating security measures that stay ahead of increasingly sophisticated adversaries. These updates to HiddenLayer's platform allow organizations to deploy AI models more securely across diverse environments while mitigating critical risks.

"It's vital that security providers keep pace with the bad actors–especially in enterprise environments, where we bear the responsibility of safeguarding our customers' most critical assets," said Chris Sestito, CEO and Co-Founder of HiddenLayer. "These new capabilities increase risk detection across the board and enable us to better serve and protect customers with more flexible and scalable options."

AISec Platform: Enterprise-Ready Security and User Management

In addition to enhanced detection capabilities, HiddenLayer's AISec Platform, which provides detection and response for AI models, is now equipped with advanced tools for managing large-scale enterprise deployments. These include comprehensive user management features and secure integration with existing enterprise infrastructure:

User Management: Enterprises can now easily manage tenant users, including creating, editing, and deleting user accounts. This capability strengthens internal control and access management across large organizations.
SAML SSO: A fully integrated Single Sign-On (SSO) and Role-Based Access Control (RBAC) experience ensures administrators can securely and efficiently assign roles and permissions. The SSO integration further enhances enterprise readiness by streamlining access for larger teams.

Enterprises are facing increased pressure to adopt AI technologies while simultaneously navigating a growing landscape of digital threats. HiddenLayer's new features allow companies to confidently scale their AI initiatives without sacrificing security or efficiency, providing a competitive edge in industries where trust and innovation are key.

"The security frameworks established by organizations like ATLAS and NIST are invaluable resources—some of which we've had the privilege to help shape. By integrating well-established security frameworks into our solutions, we're able to provide even stronger, more adaptable protection to our customers. In a world where AI plays a crucial role in day-to-day business operations, safeguarding these models is mission-critical," said Malcolm Harkins, Chief Security & Trust Officer of HiddenLayer.

Model Scanner: Increased Scalability and Risk Detection

As AI continues to become an integral part of the digital supply chain, enterprises must ensure that every component of AI-driven systems is secure from development to deployment. HiddenLayer's Model Scanner reduces the risk of adversarial attacks, with new updates offering enhanced deployment options and seamless integration into continuous integration/continuous deployment (CI/CD) pipelines.

Introducing Model Risk Context: Heightened Detection Risk Context

These updates include Model Risk Context, which enhances the depth of risk detection by mapping identified threats to widely recognized industry frameworks such as OWASP, ATLAS, and NIST. This level of visibility equips organizations with a holistic understanding of potential risks, enabling them to make informed security decisions based on the risk profile of AI models. Other updates include:

Static Analysis Results Interchange Format (SARIF): The platform now outputs SARIF from its API, allowing integration with tools like GitHub Advanced Security that support the Static Analysis Results Interchange Format (SARIF).
Local Model Scanning: Users can now conduct ad-hoc scans on local models, offering greater flexibility for proprietary or offline AI assets.
CLI Object Storage Support: This feature allows enterprises to scan models stored in AWS S3 and Azure Blob, enhancing versatility for organizations operating across multiple cloud environments.

With new integrations such as JFrog Artifactory and GitHub Actions, and the ability to scan models directly from the terminal, the Model Scanner ensures that security is embedded into every phase of AI development. Enterprises using Google Cloud Platform (GCP) can also benefit from a fully self-hosted deployment option, giving them complete control over their AI security infrastructure.

HiddenLayer's platform signals a fundamental shift in how enterprises secure their AI environments. With risk detection that maps to industry standards, seamless integration into existing workflows, and tools for flexible deployment, HiddenLayer is setting the new standard for AI security. To see how HiddenLayer's Security for AI solutions can protect your enterprise, visit the Microsoft Azure Marketplace or explore our latest Product Blog.

Media Contact: [email protected]

SOURCE HiddenLayer