Mage Data has launched a new extension for its data protection platform, aimed at safeguarding sensitive information throughout the artificial intelligence lifecycle. This latest offering, Data Security and Privacy for AI, is tailored to assist enterprises in securing data within AI training environments, public generative-AI applications, custom AI agents, and embedded copilots. The platform operates by applying data protection policies at different stages, including before data enters an AI system, during its processing and development, and when an AI system generates a response.
The challenge of applying traditional enterprise data controls to AI environments is addressed by Mage Data’s new capabilities. Sensitive information often moves through various mediums such as extracts, notebooks, feature stores, evaluation datasets, prompts, and AI-generated responses, making conventional data management difficult. Mage Data’s solution encompasses five primary areas of protection: training data guardrails, AI usage guardrails, dynamic data masking for AI, AI development guardrails, and activity monitoring for AI.
Training Data Guardrails are designed to identify sensitive information, like personally identifiable information (PII), protected health information (PHI), and non-public information (NPI), across both structured and unstructured datasets. Organizations can implement data masking at the source, protect information as it enters AI pipelines, or apply controls via software development kits. AI Usage Guardrails allow for the inspection of employee prompts and file uploads to public generative-AI services, enabling the masking of sensitive information before it exits a user’s device. Additionally, dynamic data masking can alter AI-generated responses through masking, redaction, generalization, or blocking based on user identity, request, and response content.
For organizations developing their own AI agents, AI Development Guardrails offer necessary controls, with Mage Data’s SDKs and MCP Server restricting tools and data access according to user permissions. Furthermore, Activity Monitoring for AI records interactions with AI, encompassing users, prompts, tools, sensitive data management, overrides, and policy outcomes, while also providing reporting and alerting capabilities. Mage Data emphasizes that organizations can adapt existing data policies to AI workloads without needing a separate policy framework for AI.
Mage Data’s CEO and founder, Rajesh Parthasarathy, underlines the company’s strategy of applying existing data protection principles to the various environments where enterprise data interacts with AI systems. The company also points out the potential risks associated with employees utilizing public AI tools that handle sensitive information. According to Mage Data’s CTO and Senior Vice President, Anil Bhat, their approach aims to protect data without the need for enterprises to entirely block AI tools, which could otherwise lead employees to turn to unmanaged services. The new Data Security and Privacy for AI solution is now available, with Mage Data providing demonstrations and proof-of-concept deployments for organizations interested in evaluating the technology.
Legal Disclaimer:
The information contained in this article has been provided by independent third-party contributors, clients, or content partners. We do not independently verify the accuracy, completeness, legality, ownership, licensing, or reliability of submitted content, including text, images, videos, trademarks, or other media materials. The submitting party is solely responsible for ensuring that all content, including images and media assets, complies with applicable copyright, trademark, licensing, and intellectual property laws. We disclaim liability for any unauthorized use of copyrighted or proprietary materials by third parties. If you believe that any content published on this platform infringes your intellectual property rights, kindly contact the author above for prompt review and resolution.