Rethinking Oversharing Risk and Knowledge Segmentation in the Age of AI

As large language models (LLMs) like Microsoft Copilot become embedded in enterprise workflows—from search and summarization to project tracking—they introduce a new frontier of risk: the inadvertent oversharing of sensitive information. Traditional data governance approaches, built around static controls and classification, are no longer sufficient when AI can infer sensitive insights from benign data. .Join this webcast to explore how Knostic is redefining access and identity management for the AI era with a knowledge-centric approach that emphasizes not just who has access, but who needs access. Discover how their innovative methodology—grounded in need-to-know principles, role-based knowledge segmentation, and intent-aware access policies—creates an intelligent, scalable framework for controlling AI-generated knowledge sharing.

What We’ll Discuss:

  • Simulating oversharing with LLMs through real-world prompt testing
  • Mapping implicit knowledge boundaries into enforceable policies
  • Preventing lateral "knowledge movement" across roles
  • Moving from binary access controls to intent-driven knowledge delivery

Why Register:

  • Learn from industry experts
  • Submit your questions to webcast speakers (When attending live)
  • Earn 1 CPE credit
  • Flexible access (Join live or catch the replay)

Don’t miss this opportunity to rethink your AI readiness strategy and learn how governance must evolve to keep pace with the knowledge age—turning the “department of NO” into the “department of KNOW.”

SPONSORSHIP_2023_Analyst_-_Ask_the_Expert.jpg

Thank You to Our Sponsor

knostic_logo_480.jpg

This webinar is offered free of charge through collaboration between SANS and its sponsor(s). If you prefer not to share your registration details with sponsor(s), a recorded webinar will be available approximately 30 days after its initial release through the SANS archive. To access the recording, you will need to create a SANS account, but your information will not be shared with the sponsor(s).