Mage Data extends security controls into AI workloads

Mage Data said it has launched Data Security and Privacy for AI, a platform extension that applies policy to training data, prompts, responses and agent activity across enterprise AI systems. The release is available now and is aimed at helping companies use AI without exposing sensitive data or creating a separate compliance stack.

Why it matters: - Enterprise AI creates new ways for sensitive data to move, be copied and be exposed across training pipelines, assistants and agents. - Mage Data is positioning the release as a way to apply one policy framework across databases, AI prompts, model responses and monitoring. - The approach is meant to reduce the tradeoff between AI adoption and data protection.

What happened: - Mage Data announced Data Security and Privacy for AI on August 31, 2026, in New York. - The company says the release extends its data protection platform into AI environments. - The platform applies policy to data entering an AI system, data leaving one and activity in between. - The scope covers training pipelines, public generative-AI assistants, custom agents and embedded copilots.

The details: - The release builds on Sensitive Data Discovery and Classification, Static Data Masking, Dynamic Data Masking and Activity Monitoring. - Training Data Guardrails identify PII, PHI and NPI in structured and unstructured data. - Training Data Guardrails can mask data at the source, secure it as it is provisioned into pipelines or enforce controls through SDKs. - AI Usage Guardrails inspect employee prompts and file uploads to public gen-AI tools. - AI Usage Guardrails mask sensitive content before it leaves the device. - Dynamic Data Masking for AI can mask, redact, generalize or block a response based on the user, the prompt and the response. - Dynamic Data Masking for AI is designed to protect purchased copilots and agentic apps without changing them. - AI Development Guardrails let teams scope tools and data access to end-user entitlements through Mage SDKs and an MCP Server. - Activity Monitoring for AI logs AI interactions, including who asked what, which tool was used, use of unsanctioned AI tools, what data was masked, overrides and policy outcomes. - The monitoring data feeds dashboards, end-user reports and real-time alerts. - Enterprises can extend policies already deployed on the Mage Data platform directly to AI workloads. - The company says that avoids a parallel policy framework and a separate compliance regime for AI. - The company cited recognition from KuppingerCole, Bloor Research and Forrester, along with recognition in Gartner Peer Insights and customers in financial services, healthcare and higher education.

Between the lines: - Mage Data is framing AI security as a data-governance problem, not a model problem. - The message is aimed at enterprises that want to use public AI tools and vendor copilots without giving up control over sensitive information. - The emphasis on masking rather than blocking suggests the product is designed to keep workers using approved AI tools instead of shifting to shadow IT.

What's next: - Data Security and Privacy for AI is available now. - Mage Data is asking prospects to book a demo and run a proof of concept in their own environment. - More information is available in the company's product page: More information. - Additional company information is available at Mage Data.

The bottom line: - Mage Data is extending its existing data-security stack into AI so enterprises can govern prompts, pipelines, responses and agent access under one policy set.