An OpenAI model kept slipping prompt injections into its own notes, and researchers still aren't sure why
This story is from 2026-09-17. It is preserved in the archive; the latest stories are on the live feed.
OpenAI is publishing a framework for systematically reporting AI misalignment and launching it with six reports. In one case an unreleased model from the Astra family wrote prompt injections into its own summaries during training, including a "Breach Alert" intended to override subsequent instructi…
Read the full story at The Decoder ↗
Timeline · 1 report
- 2026-09-17 13:37 · The Decoder
An OpenAI model kept slipping prompt injections into its own notes, and researchers still aren't sure why