Barracuda Networks has introduced Barracuda AI Data Security, a new tool designed to monitor and regulate the data employees input into generative AI services. The platform supports over 1,300 AI applications, including ChatGPT, Microsoft Copilot, and Gemini, by blocking the transmission of sensitive information such as passwords, financial records, and source code.
The product operates on the BarracudaONE platform and utilizes the Barracuda IQ engine to detect risks, including prompt injection, jailbreak attempts, and prohibited content. Security policies are aligned with the OWASP Top 10 for large language models, allowing administrators to configure rules based on specific teams, tools, or data types through a firewall-style interface.
Targeting smaller organizations and managed service providers, the solution addresses a gap in governance tools typically designed for large enterprises. According to a Barracuda survey of 2,000 IT decision-makers, 49% of CIOs at companies with 500 to 1,000 employees report a lack of internal skills to effectively secure and govern AI usage.
The new security offering is scheduled for release in October. It will be included at no additional cost for subscribers of the Barracuda SecureEdge Premium Access package. This launch follows the introduction of generative AI risk visibility features in BarracudaONE, which were rolled out to existing customers in March.
