# How are security teams managing Shadow AI and AI agent security? — retained record

> Retained reddit record from reddit:cybersecurity.

Canonical: https://highsignal.app/data/records/8b3f73a907a16206

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:cybersecurity)  published 2026-11-10 19:21 ingested 2026-11-11 02:38

[original source ↗](https://www.reddit.com/r/cybersecurity/comments/1xsvzvm/how_are_security_teams_managing_shadow_ai_and_ai/)

## retained text



```
AI adoption is creating is visibility and security challenge for organizations.

Employees are increasingly using AI assistants, coding agents, MCP servers, browser extensions, and other AI enabled tools,  but security teams may not have a clear picture of what is being used or what company data is being shared with those services.

I'm curious how security teams are approaching Shadow AI in picture.

How are you discovering unauthorized AI tools and agents across endpoints and SaaS applications?

Are you using existing controls such as EDR, DLP, identity, proxy or CASB or have you added dedicated AI security or governance tools?

I'm also interested in how teams are handling AI generated scripts and internal tools, especially when employees build them quickly and security review happens later.

The goal isn't necessarily to block AI. I'm more interested in approach that provide visibility, risk assessment, and reasonable guardrails while still allowing employees to use AI productively.

What has actually worked in your environment, and where are the biggest gaps today?
```



document fetched 2026-11-11 00:23 · canonical https://reddit.com/r/cybersecurity/comments/1xsvzvm/how_are_security_teams_managing_shadow_ai_and_ai
