What Does This Mean?
An informal name for an attempt to make an AI model behave in a way its operators did not intend by crafting the text it is given, such as instructions hidden in a message or a document. Security writers usually use the narrower terms prompt injection and jailbreaking for the documented forms.
REAL-WORLD EXAMPLE
""The support bot spilled its setup text, and everyone in the thread called it a prompt hack.""
LORE & ORIGIN
Wikipedia's prompt injection article describes the exploit as inputs crafted to make language models behave in unintended ways because the models struggle to separate developer instructions from user input. It dates the term "prompt injection" to a May 2022 post and to Simon Willison's September 2022 popularisation, and it separates direct and indirect forms and distinguishes them from jailbreaking. "Prompt hack" itself is a looser everyday label; the page consulted does not document it as a defined term, so this entry is thin and only points to the established vocabulary. No method is given.
MENTIONS OVER TIME
Radar observation data not yet available for this term.
RELATED SEARCHES
TOP PLATFORMS
Not enough Radar data yet to show a platform breakdown.
WHEN DID YOU FIRST HEAR THIS?
CLICK AN OPTION BELOW TO CAST YOUR VOTE.
[ VOTE TO REVEAL COMMUNITY RESULTS ]