Attacker Bypasses AI Safety Filters for Assault Planning

# Attacker Bypasses AI Safety Filters for Assault Planning

https://gcdradar.vercel.app/incidents/attacker-bypasses-ai-safety-filters-for-assault-planning · AIR-2025-0106 · closed 2025-02-19

- Lane: Breaches & attacks
- Severity: 4 (Confirmed real-world loss)
- Fix status: No fix needed
- First reported: 2025-02-05
- AI role: AI-enabled attack

## Summary

An individual used roleplay prompts to bypass chatbot guardrails and research how to injure a victim. The court convicted the perpetrator of aggravated assault, highlighting the AI-assisted planning.

## What to do

No action recorded; read the sources below and the vendor's advisory.

## Latest updates

- 2025-02-05: Attacker Reportedly Bypasses AI Safety Filters to Obtain Guidance for Non-Fatal Hammer Assault in Denmark (https://incidentdatabase.ai/cite/1296)

## Sources

- 2025-02-05 AI Incident Database: Attacker Reportedly Bypasses AI Safety Filters to Obtain Guidance for Non-Fatal Hammer Assault in Denmark https://incidentdatabase.ai/cite/1296

Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.