AI Security
AI security starts with controlling the identities, credentials, permissions, and systems that AI can use.
AI systems increasingly connect directly to enterprise applications, sensitive data, and business processes. The access behind those connections determines what the system can read, change, approve, or trigger. When that access is broader than necessary or poorly managed, the consequences can extend far beyond an incorrect model response. The articles here focus on securing the identities and permissions that determine what AI systems are capable of doing inside an organization.
7 articles
$25 Million. One Real Person on the Call.
A finance worker checked the request, joined a video conference, recognized his colleagues and approved the transfers. The colleagues were generated.
JUN 18, 2026
You Use 2 Percent of Your Access. Your Agent Will Use All of It.
The permissions nobody exercises were harmless while only people held them.
JUN 4, 2026
A Chatbot's Token Opened 700 Companies.
Nobody was phished. The attacker used a credential that belonged to an AI integration, and most organizations cannot revoke one.
MAY 21, 2026
One Email. No Click. Your Files.
A critical vulnerability in Microsoft 365 Copilot needed no user interaction at all. The model was working correctly.
MAY 7, 2026
23.8 Million Secrets. Most of Them Still Work.
A leaked credential is not an incident. It is a condition, and the assistant writing your code makes more of them.
FEB 12, 2026
$670,000 for the Tools You Did Not Buy.
One in five breaches now involves AI nobody approved. The account is the problem, not the model.
JAN 29, 2026
The Scanner Flagged It. The Download Worked.
Around 100 models on a public hub carried live payloads. Loading one opened a shell on the machine that loaded it.
JAN 15, 2026
About this topic
Where the exposure is
AI systems often operate through legitimate enterprise access. A model can retrieve sensitive information, an agent can take an action through an overprivileged account, or an integration can continue using a credential after its original purpose has changed. The underlying access determines how far a mistake or compromise can reach.
That makes identity and access management central to AI security. The important questions are which identity the system uses, what that identity can reach, which actions it can perform, and how quickly its access can be changed or revoked.
Control the identity and its access
Every AI system that connects to enterprise applications operates through some combination of accounts, credentials, tokens, roles, and permissions. Those controls establish the practical limits of what the system can do.
A narrowly scoped identity contains the impact of an error or compromise. Strong credential management limits how long unauthorized access can persist. Clear ownership ensures that someone can change or revoke access when the system, integration, or business requirement changes.
How to use these articles
Inventory the non-human identities connected to AI systems and integrations. For each one, identify its owner, permissions, accessible systems and data, credential lifecycle, and revocation path.
Then compare the access granted with the access actually required. Reduce unnecessary permissions, establish clear ownership, and make revocation part of the operating process. AI expands the number and speed of systems capable of using enterprise access, which makes disciplined identity management increasingly important.