Mage Data has unveiled a new feature called Data Security and Privacy for AI, which enhances its existing data protection platform to help businesses safeguard sensitive information when utilizing artificial intelligence. These new capabilities are tailored to secure data throughout the AI lifecycle, covering areas such as AI training environments, public generative-AI applications, custom AI agents, and embedded copilots. The platform ensures that data protection policies are enforced before data is fed into an AI system, during its processing and development, and when AI outputs are generated.
The challenge of applying traditional enterprise data controls to AI environments stems from the movement of sensitive information through various stages such as extracts, notebooks, feature stores, evaluation datasets, prompts, and AI-generated outputs. Addressing this issue, Mage Data’s new solution offers five key protection areas. Training Data Guardrails can identify sensitive data types, including personally identifiable information (PII), protected health information (PHI), and non-public information (NPI) within both structured and unstructured datasets. This allows organizations to mask data at its origin, safeguard it as it enters AI pipelines, or impose controls via software development kits.
AI Usage Guardrails are designed to monitor employee interactions with public generative-AI services, ensuring that sensitive information is masked before leaving a user’s device. Dynamic Data Masking for AI allows for the masking, redaction, generalization, or blocking of AI responses based on the user, the request, and the information contained in the response. Additionally, AI Development Guardrails provide necessary controls for organizations developing their AI agents, with Mage Data’s SDKs and MCP Server limiting tool and data access according to user permissions.
To further enhance security, the platform includes Activity Monitoring for AI, which logs AI interactions, including user prompts, tools, and sensitive data masking, along with any overrides and policy outcomes, while offering robust reporting and alerting features. Mage Data emphasizes that existing data protection principles can be seamlessly extended to AI workloads, eliminating the need for a separate policy framework specifically for AI. According to CEO and founder Rajesh Parthasarathy, this approach is key to managing the interaction between enterprise information and AI systems.
The risk of employees inadvertently using public AI tools with sensitive data is a significant concern. Anil Bhat, the CTO and Senior Vice President of Mage Data, notes that their strategy is crafted to protect data without necessitating a complete ban on AI tools, which could otherwise drive employees towards unregulated services. With Data Security and Privacy for AI now available, Mage Data is offering demonstrations and proof-of-concept deployments for organizations interested in evaluating this 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.
