The cybersecurity industry has spent decades building walls against external attackers but has quietly handed the keys of every wall to an AI assistant that has access to emails, calendars, codebases and customer data without most organizations fully understanding what happens when that assistant is manipulated, because the most dangerous security vulnerability in 2026 is not a hacker trying to brute force a password but a trusted AI agent that can be redirected by a single cleverly crafted prompt hidden inside a document or an email it was simply asked to summarize. The most expensive breach of the next decade will probably not come from outside the organization but from inside a workflow that everyone approved and nobody fully understood.
-
Technical Discussion
Tech talk
-
Announcements
Announcements regarding our community
-
General Discussion
A place to talk about whatever you want