Table showing four categories of AI data leakage: Application-level (leakage through AI app responses, e.g. LLM outputting another user's data due to context contamination), User-introduced (data entered by employees, e.g. pasting customer PII into ChatGPT), Model/training data (memorized data extracted from models, e.g. extraction attacks revealing training set contents), and Agent/workload (autonomous systems leaking via integrations, e.g. an agent with database access returning unauthorized records.

What Causes AI Data Leakage and Tips for Staying Protected

9 min readWhat it takes to implement it, and why real-world environments make it hard to finish. The post What Causes AI Data Leakage and Tips for Staying Protected appeared first on ...
↗

Anthropic Workload Identity Federation: What It Gets Right – and What It Still Doesn’t Solve

8 min readWhat it takes to implement it, and why real-world environments make it hard to finish. The post Anthropic Workload Identity Federation: What It Gets Right – and What It Still ...

AIMS: A Model for AI Agent Identity

4 min readAI agents are starting to look less like tools and more like participants. They don’t just respond to prompts. They take actions, call APIs, orchestrate workflows, and, in some cases, ...

Gartner’s Workload IAM Architecture Is a Big Step Forward for AI Agent Security

8 min readAI agents are often described as a new identity problem. That is true, but it is only half the story. Agents do introduce new behaviors. They can dynamically choose tools, ...

How a Long-Lived API Credential Let an AI Agent Delete Production Data

4 min readWhat began as a routine staging task for a SaaS startup ended in a disaster that  would have been unthinkable just months ago: an AI agent operating as a super ...
AI listing of Aembit workloads with active access policies shown.

Analyze AI Agent Access: Introducing the Aembit MCP Server

4 min readTeams can query workload identity data in plain language, investigate activity, and move faster without leaving the Aembit platform. The post Analyze AI Agent Access: Introducing the Aembit MCP Server ...
[Demo] Aembit IAM for Agentic AI

Aembit IAM for Agentic AI Is Now Generally Available

5 min readAembit IAM for Agentic AI is now GA. Enforce AI agent access to MCP servers with Blended Identity, secretless credential exchange, and policy-based enforcement. Free tier is available. The post ...
Aembit logo

MCP or CLI? How to Choose Right Interface for Your AI Tools

4 min readWhat starts as a tooling decision ends up shaping cost, reliability, and how far your workflows actually scale before they break down. The post MCP or CLI? How to Choose ...
Aembit and Netskope architecture

Announcing the Aembit + Netskope Partnership for Agentic AI Security

5 min readOAuth is an authorization framework that defines how to grant access. JWT is a token format that defines how to package and transmit claims. They solve different problems, and most ...

The OWASP Top 10 for LLM Applications (2025): Explained Simply

6 min readThe OWASP Top 10 for LLM Applications is the most widely referenced framework for understanding these risks. First released in 2023, OWASP updated the list in late 2024 to reflect ...