QuestionQ55

VerifiedIdentify benefits, capabilities, and opportunities for Microsoft’s AI apps and services

For each of the following statements, select Yes if the statement is true. Otherwise, select No.

Yes or No
StatementsYesNo
Content filtering controls can prevent AI-generated responses from exposing confidential and sensitive information.
AI-generated content can unintentionally reveal sensitive information if the generative AI model has access to unsecured data sources.
To prevent data exposure, only the prompts used by users must be protected by using policies.
Explanation

Content filtering mechanisms such as Azure AI Content Safety are built to detect and block harmful content categories (hate, violence, sexual content, self-harm) in prompts and completions; they are not a substitute for data loss prevention and are not designed to stop confidential or sensitive business data from being exposed in AI responses — that requires DLP/sensitivity-label based controls such as Microsoft Purview policies for Copilot and other generative AI apps. When a generative AI model has access to unsecured or ungoverned data sources (e.g., files or repositories without proper permissions/sensitivity labels), it can retrieve and unintentionally surface confidential information in its generated output, which is why data security posture management and access controls on data sources are required. Finally, protecting only user prompts is insufficient — effective data-exposure prevention policies (as implemented through Microsoft Purview DLP for Microsoft 365 Copilot and other generative AI apps) must govern both the prompts sent to the model and the AI-generated responses/interactions returned to users.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!