Cloudflare has launched new AI security features to help organizations safeguard against AI-related threats, including a Shadow AI Report that tracks AI application usage and users.

The platform now enables automatic enforcement of AI policies at the network edge via Cloudflare Gateway, allowing security teams to block unapproved AI tools, control data uploads, and maintain compliance.

These tools include AI Security Posture Management for compliance, the Shadow AI Report for monitoring AI app activity, and policy enforcement through Cloudflare Gateway.

The new features aim to mitigate risks associated with generative AI use by employees, such as data sharing and deploying unapproved AI applications.

As AI adoption increases, managing AI risks is crucial to maintaining security and privacy standards across organizations.

Cloudflare's Zero Trust MCP Server Control consolidates AI model interactions and external tool traffic into a single dashboard, providing better access management and policy enforcement.

This centralized dashboard offers visibility and control over all MCP tool interactions, enabling security teams to enforce policies at both user and server levels.

These new capabilities allow organizations to understand, analyze, and control AI usage comprehensively, promoting productivity and innovation without compromising security or privacy.

Organizations can monitor AI activity, prevent Shadow AI, and protect sensitive data from a centralized platform, enhancing overall security posture.

Widespread AI adoption often occurs without proper security measures, risking data leaks and unauthorized AI deployment.

Cloudflare announced enhanced security features for its Cloudflare One zero-trust platform to facilitate secure adoption and deployment of generative AI applications.

CEO Matthew Prince highlighted Cloudflare’s unique ability to combine zero-trust security with AI development tools, supporting secure innovation.

AI Prompt Protection is introduced to detect and mitigate risky employee interactions with AI models by flagging or blocking sensitive prompts and responses.

This feature monitors prompts in real-time to prevent data leakage and manage risky interactions with AI models.

Key security capabilities include AI Security Posture Management, Shadow AI Report, AI Prompt Protection, and Zero Trust MCP Server Control, which collectively enhance AI activity visibility and control.

Cloudflare is also deploying Zero Trust MCP Server Control to manage AI model interactions with external systems by routing traffic through Cloudflare for inspection.