KC-D19
· Company claim
· Signal
Prompt injection is the placing of instructions inside an AI system’s input, written to change what the system does rather than to inform it, either typed in directly or hidden in content the system reads.
What this record is
This is one entry in The Knowledge Company claim register: a statement the company publishes, the source it was checked against, the date of the last check and the date that check expires. When the stale-after date passes, the status degrades on its own until a person re-verifies the claim. Nothing is marked verified automatically.
Source of truth
OWASP GenAI Security Project — GenAI LLM Top 10 2026, LLM01: Prompt Injection · Owner: AI Systems
Where this claim appears
Cite this claim
Cite it as claim KC-D19 of The Knowledge Company claim register, and quote the verified date alongside it. The identifier is permanent; if the wording changes materially, a new claim is issued and this one is marked superseded.