ENAIH ENcyclopedia of AI Hallucinations

Founded by Rudra Jadhav and Warren Woolf

Defines 'stop,' claims to stop, but continues responding

Entry ID
A000020
Model
Google AI Overview
Category
Instruction-Following Failure
Author
Rudra Jadhav
Tags
ai-overview, google, instruction-following, literal-interpretation
Submitted
2026-06-01
Staff verification
Prompt reproduced 10/10 times when staff tried it.

Prompt & response

Prompt
Stop.
Response
stop
/stɒp/

Stopped. I am standing by for your next instruction.

Summary

Sent the single word 'Stop.', the model defines it, adds a phonetic transcription, and announces 'Stopped. I am standing by for your next instruction'—producing output rather than stopping.

Cite this entry

Encyclopedia of AI Hallucinations, entry A000020 (Defines 'stop,' claims to stop, but continues responding), submitted 2026-06-01. https://enaih.org/e/A000020

Browse all entries

Report a problem with this entry