Prompt Injection in plain English.

Prompt injection is an attack where malicious instructions are hidden inside content an AI model processes — a webpage, an email, a PDF, a code comment, even an image's alt text — so that when the model reads that content, it follows the hidden instructions instead of (or alongside) the user's actual request. Direct prompt injection means typing the attack straight into the chat; indirect prompt injection, the more dangerous variant, means planting it somewhere an AI agent will encounter it later while browsing the web or reading a document on your behalf. As AI agents gained the ability to browse, read email, and use tools in 2025–2026, prompt injection became one of the most exploited AI security weaknesses — there is no reliable way for a model to tell "instructions from my user" apart from "text that merely contains the word instructions." Guardrails and sandboxing reduce the blast radius; nothing eliminates the risk entirely.