GCD AI RADAR
incidents/AIR-2026-1400

ECHO: Response Harmfulness Optimization-Based Jailbreak Attack

Attack on AI1 Research / proof of conceptJailbreak

Researchers propose ECHO, an optimization-based jailbreak attack that replaces compliance proxies with direct response harmfulness evaluations. The method demonstrates improved capability in bypassing model safety alignments.

Details

type
Attack on AI
An AI model, agent or AI-powered product was the thing exploited, abused or leaked.
categories
  • Jailbreak
    A user deliberately bypasses a model's safety or policy controls through crafted prompts.
owasp llm top 10
mitre atlas
Both rows are derived from the type and categories (OWASP 2025 edition, ATLAS release 2026.09), not classified per record. How
severity
1 · Research / proof of concept
Demonstrated by researchers; no known real-world exploitation.
targets
Open-weight model
products
ECHO
incident date
19 Oct 2026
first reported
19 Oct 2026
origin
News & research feeds
classifier confidence
0.90 · published at or above 0.80; see the methodology
tags
jailbreak · optimization · safety-alignment
record
created 2026-10-19 05:07Z · updated 2026-10-19 05:07Z · json

Sources · 1

  1. 2026-10-19
    ECHO: Replacing the Compliance Proxy with Response Harmfulness in Optimization-Based Jailbreak Attack
    arXiv · arxiv.org/abs/2610.23030

Titles and links belong to their publishers. The summary above is written by GCD AI RADAR in its own words; data is CC BY-SA 4.0.

Related incidents