SECSEC-001 — Model Security Vulnerabilities
Adversarial Prompt Manipulation Extracts Restricted LLM Outputs
4/5Sector: OtherGeography: GlobalStage: OperateIngested: —
Executive Summary
Controlled prompt perturbations can reverse GPT classification decisions and bypass content refusals to extract dangerous information. Firms deploying LLMs in regulated workflows face material liability where adversarial inputs circumvent compliance controls.
Domain
Security & Privacy
Blindspots in model security, data poisoning, privacy leakage, infrastructure, model theft, and incident response.
Source
MIT AI Risk Repository — Trustworthy LLMs: A Survey and Guideline for Evaluating Large Language Models’ Alignment (Liu2024) ↗https://airisk.mit.edu/
Could this happen in your organisation?
A Velinor AI Audit maps your active AI portfolio against the 50+ blindspots and benchmarks against documented sector failures like this one. A board-ready foresight document in 5 weeks.