SECAI Core · phase 11 of 15
Obfuscation
A technique that essentially 'hides' or 'camouflages' code or other information so that it is harder to read by unauthorized users.
The Explain card
- Plain English
- Obfuscation hides or camouflages code or data so it is harder for unauthorized readers to understand, without changing what it does.
- Example
- A prompt injection is written in base64, spread across several invisible HTML spans, or translated into an unusual language, so a simple keyword filter never sees "ignore previous instructions" while the model still decodes it.
- Why it matters
- Attackers obfuscate to slip past filters and analysts; defenders sometimes obfuscate to slow reverse engineering. For AI, input filters that only match plain text are easy to sidestep.
- Hook
- Same message, different disguise.
Where it sits in the deck
Phase 11: Threat Landscape: Attack Vectors and Adversarial Techniques
With defences named, learn what they must defend against — the full catalogue of attack techniques targeting AI systems, their inputs, outputs, training pipelines, and supply chains.
