Definition

Prompt Injection

Prompt Injection is a security vulnerability where a malicious user provides input that overrides the pre-configured system instructions or safety alignment filters of a Large Language Model, hijacking its control flow.

Frequently Asked Questions

What is the difference between direct and indirect prompt injection?▼

Direct prompt injection occurs when a user prompts the model to ignore rules. Indirect prompt injection happens when a model reads untrusted external data (like a webpage) containing malicious instructions.

How do you defend against prompt injection?▼

By separating user input from developer instructions, using strict system prompt isolation, and employing secondary filter models to inspect outputs.

Quick Facts

CategoryModel Limitations
Key ApplicationVulnerability analysis, LLM firewall design, and input sanitation.

Coverage Trend12 Weeks

12w agoToday

Related AI Terms

Adversarial Attack AI Safety Guardrails

Prompt Injection Media Coverage & Intelligence

TechCrunch AIJun 6, 2026

OpenAI unveils Lockdown Mode to protect sensitive data from prompt injection attacks

Even with Lockdown Mode, ChatGPT could be vulnerable to prompt injections, but the goal is to reduce the likelihood that sensitive data gets shared in the proce

Read Original Coverage