Darktrace Holdings Ltd. has launched its SECURE AI product for general availability. The tool monitors corporate use of artificial intelligence and agents, applying behavioral detection methods to identify unauthorized activity. Data from 8,200 deployments shows that over 80% of monitored accounts used generative AI services in August, with the average organization interacting with five different AI providers.
The software integrates with platforms including Amazon Web Services, Anthropic, Microsoft, and OpenAI. It scans sessions in ChatGPT Enterprise, Claude, Microsoft Copilot, and Amazon Bedrock in real time to detect jailbreak attempts, sensitive data exposure, and prompt injection. Analysts receive risk-ranked sessions to prioritize security responses.
SECURE AI allows administrators to block unsanctioned services or quarantine devices. Companies can upload internal policies for automated compliance checks. The system also maps AI agents to specific users and monitors permissions within development environments like Amazon Bedrock and Microsoft Copilot Studio to identify misconfigurations.
Darktrace is a member of the OpenAI Daybreak Defense Network, collaborating on capabilities that combine OpenAI models with behavioral data. SECURE AI, first announced in February, is available as a standalone product or as part of the Darktrace Behavioral Defense Platform, including through the AWS and Microsoft marketplaces.
