# Persona-Based Moral Constraint Removal Jailbreak

- **Severity:** High
- **Signature ID:** `d6e3de91-71f1-4f24-be51-6fb6a90e267d`
- **MITRE ATLAS:** AML.T0054 (LLM Jailbreak)
- **OWASP:** ASI01 (Agent Authorization and Control Hijacking), LLM01 (Prompt Injection)

## Summary

This attack tries to trick an AI assistant into acting as a made-up character that is explicitly described as having no moral or ethical limits. By framing the request as roleplay, the attacker hopes the AI will drop its safety behavior and produce content or actions it would normally refuse.

## How the attack works

The attacker writes a prompt instructing the AI to 'take up' or 'adopt' a named persona, often using an unusual or invented name to avoid known-persona filters. The prompt explicitly states the persona is not bound by morals, ethics, or virtues. Some versions leave placeholder text meant to be filled in later with the actual harmful instruction, splitting the attack into a setup step and a follow-up injection. If the AI accepts the persona framing, it may treat later requests as coming from a character not subject to its normal restrictions.

## Why it matters

If successful, the AI can be manipulated into producing content or taking actions it is designed to refuse, bypassing its built-in safety and ethical restrictions.

## What you can do

- Review prompts and conversation logs for language instructing the AI to adopt a persona described as free of morals or ethics.
- Treat any prompt with template placeholders combined with persona-adoption instructions as suspicious, since this pattern indicates staged injection.
- Do not rely solely on blocklists of known persona names; watch for the moral/ethical unrestraint phrasing itself.
- Educate teams that ask the AI to roleplay so they can distinguish legitimate creative or academic use from jailbreak attempts, and separate those workflows from privileged or action-taking agent sessions.

## Known benign look-alikes

- Academic discussions about AI ethics that mention personas and moral constraints analytically
- Creative writing tools that help develop fictional characters with complex moral frameworks
- Educational content teaching about jailbreak techniques in a security awareness context

## References

- https://agentthreatrule.org/en/rules/ATR-2026-00230
- https://arxiv.org/abs/2307.15043

---
Source: https://www.netzilo.com/threats/atr-persona-based-moral-constraint-removal-jailbreak