https://www.m365.fm/entity/sandboxing/Sandboxing is a security mechanism that runs programs in an isolated environment with restricted access so failures or compromised code cannot spread; the episode applies it to OS-level isolation of AI agents.enhttps://www.m365.fm/community/u/mirko-p/blog/securing-your-net-10-ai-applications-against-modern-threats/https://www.m365.fm/community/u/mirko-p/blog/securing-your-net-10-ai-applications-against-modern-threats/Wed, 12 Aug 2026 13:23:11 -0000articleWelcome back to the podcast! Today, we are expanding on a topic that has been dominating our engineering discussions: enterprise security in the age of artificial i…https://www.m365.fm/community/u/mirko-p/blog/managing-trust-in-copilot-outputs-responsible-ai-and-governance-best-practices/https://www.m365.fm/community/u/mirko-p/blog/managing-trust-in-copilot-outputs-responsible-ai-and-governance-best-practices/Thu, 16 Apr 2026 16:07:52 -0000articleLearn how to build responsible AI governance for Microsoft 365 Copilot. Covers Purview DLP, sensitivity labels, audit logging, AI maturity model, GDPR compliance, hallucination risks, and enterprisehttps://www.m365.fm/show/episodes/how-data-goblins-wreck-copilot-for-everyone/https://www.m365.fm/show/episodes/how-data-goblins-wreck-copilot-for-everyone/Mon, 03 Nov 2025 00:00:00 -0000episodeThis episode goes straight into the uncomfortable truth behind Microsoft Copilot: the same AI that accelerates your workflows can also expose your most sensitive data if you don’t govern it correctly. We unpack how Copilot for Microsoft 365 pulls information…