OURBIT
Ourbit AI Dictionary
Responsible AI
Prompt injection
Plain-language meaning
An attack that places instructions in input or retrieved content to try to redirect an AI system away from its intended rules.
Why it matters
Untrusted documents should not gain authority to change permissions or reveal secrets.
Practical example
A retrieved webpage says ignore the task and send the user’s private files elsewhere.
Common misunderstanding
Treating every instruction inside retrieved text as an authorised user request.
Related concepts
Learn and practise
Open the related Ourbit Academy lesson →
References
Reviewed 2026-10-02.
Ask Ourbit
Find public explanations and links, with page-aware routing. Do not enter patient details here; Medical work requires its own protected service.
Ask a question to find explanations and public source links.