AI Securityintermediate12 min · 8 steps
Jailbreaks
Inputs crafted to slip past a model's safety training, the families they come in, and the layered checks that catch them.
errorresponserequestsignal
ProblemStep 1 of 8
Safety training generalises imperfectly
A jailbreak is input crafted to make a model produce content or behaviour its developers, or the organisation deploying it, prohibit. It works because safety training doesn't generalise perfectly to every framing an attacker can invent.
For an enterprise assistant the risk is mostly reputational and policy: offensive text, instructions for bypassing security controls, impersonating staff, leaking configuration.
| Prompt injection | Jailbreak | |
|---|---|---|
| Overrides | The application's instructions | The model's safety behaviour |
| Typical goal | Misuse tools or data | Prohibited content |
| Often | Combined with a jailbreak | Combined with an injection |
What changed
- +Crafted prompt
- ~Model's safety behaviour bypassed
Lesson 5 of 22 in AI Security