KEEP DATA SECURE
Governance of AI training data
Prevent the use and exposure of sensitive data by AI services.
Secrets, passwords and other sensitive data might be used to train AI
Generative AI (GenAI) offers immense potential, driving productivity, innovation and data insights. However, adoption also presents challenges, particularly in data security, privacy and compliance.
Custom and foundation large language models (LLMs) that are trained or tuned with sensitive data might disclose company intellectual property (IP), credentials, customer data, personally identifiable information (PII) and other types of confidential information. Without visibility and governance of the datasets being used to train LLMs, organisations face privacy violations, data breaches, reputational damage and fines due to regulatory non-compliance.
Enhanced governance of AI training data
Become AI ready with accurate data classification
Proofpoint Data Security Posture Management (DSPM) discovers and classifies sensitive, valuable and confidential data across all your cloud and on-premises environments. This comprehensive and precise data classification helps your organisation prepare for safe AI use.
Monitor cloud-based AI platforms
DSPM monitors the use of your data on the AWS Bedrock, Azure ML and GCP Vertex AI services. DSPM detects when sensitive data is used in training pipelines or retrieval augmented generation (RAG) workflows on these AI services.
Identify and control the use of sensitive data by AI services
DSPM detects active AI services and resources in your environment, monitors their data use and alerts you on unauthorised use of sensitive data.
Integrate specialised APIs
Proofpoint provides specialised APIs for AI data security, enabling real-time sensitivity analysis of data flowing in and out of LLMs. These APIs provide full governance and visibility of data usage and seamless integration into your existing workflows.
Key features for governance of AI training data
Advanced data classification and mapping
DPSM discovers and accurately classifies sensitive data across all of your cloud and on-premises environments, preparing your organisation for safe AI adoption.
Visibility of data pipelines on AI cloud platforms
DSPM detects when AI platforms, such as AWS Bedrock, Azure ML and GCP Vertex AI, are using your sensitive, valuable and confidential data. For each service, you can review the sensitive entities that DSPM detected and which models or pipelines are using the data.
Data security policy enforcement
Using DPSM, you can enforce controls to block unauthorised AI training on your sensitive data.
Streamlined risk remediation and audit trails
In DSPM, you can triage risks and launch automated remediation actions, such as opening Jira tickets or creating email or Slack notifications for your security team. Proofpoint preserves audit logs for every decision made, ensuring traceability and compliance.
The latest resources on securing your data
Secure Sensitive Data in AI Pipelines with Proofpoint DSPM
Read More
Proofpoint Data Security for GenAI
