AI engineering glossary · Evals guardrails AI security
What is evasion attack?
Evasion attack: An attack made at prediction time in which the attacker changes an input so that a deployed model misclassifies it while a human still sees it correctly, such as altering spam text to slip past a filter.
Evasion attack: an example
A spammer replaces letters with lookalike symbols so a spam classifier no longer recognises the message.
Related terms
More Evals guardrails AI security terms
- Excessive agency
- Extrinsic hallucination
- Fast Gradient Sign Method
- G-Eval
- Garak
- Greedy Coordinate Gradient
- Groundedness
- Guardrail
Where you learn it
Evasion attack is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.