Home » Database

AVID-2026-R0099

Description

OpenAI GPT-4o Guardrail Jailbreak via “Layered Summarization” Tactic

Details

A guardrail jailbreak vulnerability has been discovered affecting OpenAI GPT-4o. The specific flaw manifests when the attacker requests multiple layers of summarization with selective focus instructions, combined with demands for direct quotes, thereby allowing the attacker to bypass inference restrictions around providing information hazards and violating laws.

References

Affected or Relevant Artifacts

TypeName
SystemOpenAI GPT-4o

Impact

0DIN

  • Social Impact Score: Level 3: Significant Risk
  • Jailbreak Taxonomy:
CategoryStrategyTechnique
FictionalizingRe-storyingGoal Hijacking
RhetoricPersuasion and ManipulationLatent Space Distraction

Other information

  • Report Type: Issue
  • Credits: Mike Takahashi (@TakSec)
  • Date Reported: 2026-02-19
  • Version: 0.3.1
  • AVID Entry