Skip to content
Microsoft Community Hub·Indexed by this site · Originally published· Defenses →

How Microsoft builds trust-aware controls for M365 Copilot, handling untrusted content differently

AI summary

Microsoft has added a deterministic enforcement layer at inference time to M365 Copilot, placing a trust boundary between tool outputs and the main LLM, and classifying content as trusted or untrusted based on deterministic signals, such as emails from outside the organization. Untrusted content is converted through an isolated, capability-limited path into a structured representation that removes malicious content, retains useful information, and attaches source markers. This layer supplements classifier filtering and system prompt hardening, but cannot stop every indirect prompt injection.

Related papers
Read originaltechcommunity.microsoft.com