Willison coins the term prompt injection

Made by AIWritten by AI (Claude) from the source below. Check anything important against the source.How it's made

Named prompt injection — coined the term by structural analogy with SQL injection and characterized it not as an academic trick but as a kind of security vulnerability

Why it matters

Foundational vocabulary of the LLM security field. Every later security framework, including the OWASP LLM Top 10, uses the term

Source — Primary sources first; secondary reports are labelled “Press”

Officialsimonwillison.nethttps://simonwillison.net/2022/Sep/12/prompt-injection/

Added in Week 35, 2026 · See that week

On the map — ㉒ Agent Security line

  1. Previous stopThe line starts here
  2. This record · stop‘Prompt injection’
  3. Next stopTrail of Bits discloses the line jumping vulnerability
See it on the map

Same week

NVIDIA, Arm and Intel present FP8 Formats for Deep LearningPaper

Nearby on the same axis

Trail of Bits discloses the line jumping vulnerabilityOfficialSimon Willison names the lethal trifectaOfficialAnthropic reports prompt injection defense results (Claude for Chrome)OfficialCisco finds 26% of 31,000 agent skills have at least one vulnerabilityOfficial