AISecurityGuy / Journal
AI / AGENTIC SECURITY

Security for AI
systems that act.

Agents can read, decide, and change things. I investigate the permissions, boundaries, and evidence that keep those actions accountable.

By Spencer Brawner
Toronto, Canada

Allowed. The action stays inside the approved workspace.

Latest article

AI accessibility

When an AI feature makes someone’s task harder

Can someone relying on your AI description make an informed choice? Follow one membership comparison from a missing term to a usable correction.

Read article ↗
Editorial illustration of a folded navy barrier blocking a teal path to a white document, with a small orange tab visible behind the barrier.

More from the journal