Defense in Depth — AI Security Poster

Security isn't a wall — it's an onion. Defense in Depth shows the five layers that wrap an autonomous AI agent, outside in: Instruction, Output, Action, Network, Environment. Each stops a different class of attack, rendered as concentric rings closing around the agent at the core. The attacker has to beat every ring; you only need one to hold.

The insight most posters miss and this one makes plain: traditional security guards access — AI security must also guard what the model generates and does, beca

• Five defense layers, outside in — Instructiwork · Environment

• Each ring stops a different class of attack; beat them all, or don't get in

• Modeled on defense-in-depth doctrine, built

• Museum-grade matte, heavyweight archival stock, rich blacks, crisp technical type

• Printed on demand & shipped worldwide — madte

Model & Machine — All things AI.