Darktrace launches SECURE AI to oversee enterprise AI

Darktrace launched Darktrace / SECURE Artificial Intelligence (AI), a behavioral AI security product designed to give organizations visibility into how AI systems behaved, interacted with people and systems, and where risks emerged.

Darktrace released accompanying research showing more than three-quarters of cybersecurity professionals were concerned about AI agents and third-party Generative AI (GenAI) tools; the company said more than 70% of its customers used GenAI tools and that among customers with a dominant GenAI tool, 91% had employees using additional AI services. Darktrace observed anomalous GenAI data uploads averaging 75MB (about 4,700 pages) per account over five months, with some accounts averaging over 200,000 pages.

Darktrace / SECURE AI applied the company’s behavioral AI approach to monitor GenAI and agentic workflows, analyzing prompt language and data access patterns to detect anomalous activity and to enable intervention when systems drifted, exceeded authorized access, violated policy, or appeared manipulated.

The product integrated with Security Operations (SecOps) and with Darktrace ActiveAI Security Platform customers and was designed for enterprises operating AI across embedded Software-as-a-Service (SaaS) features, cloud-hosted models, and autonomous or semi-autonomous agents developed in low- and high-code environments; it monitored prompts, sessions, and model responses in tools such as ChatGPT Enterprise and Microsoft Copilot and helped track agents, Marketing Automation Platform (MAP) access permissions, and discover Shadow AI.

“Security has always been about behavior,” said Jack Stockdale, Chief Technology Officer at Darktrace. “As AI becomes agentic, prompts become the behavioral layer, encoding intent, context, and downstream actions. If you can’t observe and understand prompt language at runtime, you can’t detect drift, misuse, or emergent behavior. Securing AI without prompt visibility is like securing email without reading the message body. Prompts are to AI what traffic is to networks and identity is to users.”

Darktrace also released the 2026 State of AI Cybersecurity Report and listed additional resources, including a launch event and a Secure AI Readiness Program.