GCD AI RADAR
incidents/AIR-2026-0397

Hidden Prompt Injection Targeted Brazilian Labor Court AI Galileu

Attack on AI3 Confirmed exploitationPrompt injection

A Brazilian labor court AI tool detected and blocked hidden instructions embedded within a legal petition attempting to alter its processing behavior. The judge reviewed the material and prevented any malicious procedural outcomes.

Details

type
Attack on AI
An AI model, agent or AI-powered product was the thing exploited, abused or leaked.
categories
  • Prompt injection
    Untrusted content (web pages, documents, emails, tool output) steers a model into ignoring its instructions.
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
3 · Confirmed exploitation
Used against real targets; limited or unclear impact.
targets
Enterprise deployment
products
Galieu
incident date
13 May 2026
first reported
13 May 2026
origin
AI Incident Database
Derived from the AI Incident Database (CC BY-SA 4.0).
classifier confidence
0.95 · published at or above 0.80; see the methodology
tags
prompt_injection · legal_ai
record
created 2026-10-18 06:33Z · updated 2026-10-18 06:59Z · json

Sources · 1

  1. 2026-05-13
    Hidden Prompt Injection in Brazilian Labor-Court Petition Reportedly Tried to Manipulate Galileu
    AI Incident Database · incidentdatabase.ai/cite/1497

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