Papers & Research

OpenAI reports self-replicating prompt injection

Original title:Self-replicating prompt injections exist

AI summary

OpenAI demonstrates a new type of prompt injection that can self-propagate like a worm, discovered on June 27, 2026, based on GPT-Red self-play red-team training, with the attack model based on GPT-5.4-mini and tested models including GPT-5.4-mini and GPT-5.5.

Key Takeaways

  • OpenAI demonstrates a new type of prompt injection that can self-propagate like a worm,discovered…
  • 4-mini and tested models including GPT-5.
  • 4-mini and GPT-5.

Quick Facts

Published
Updated
Source
OpenAI:失准报告与通报(网页)
Category
Papers & Research
AI score
69
Read the original Read the full article at the original source
Why it matters

OpenAI 披露自复制提示词注入的实测案例与防御计划,展示了攻击如何借邮件和文件系统传播。

Related entities

Source & copyright

This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.

Original article · AIHOT entry

How to Cite

When citing this article, keep the title, date, and permalink, and credit Orange News as the source.

Orange News: OpenAI reports self-replicating prompt injection. September 25, 2026. https://new.nat6.net/en/news/l0p27y7t9kdxf7egvf32bob3w/

View all →