# @aiupdates.hub on Instagram

- **Type:** Image
- **Original URL:** https://www.instagram.com/p/DdmIJ8KEfHB
- **Gondola URL:** https://gondola.cc/posts/70804712-aiupdateshub-instagram
- **Thumbnail:** https://img.gondola.cc/tr:w-,h-,fo-auto/postThumbnails/2a2d3bc6f7.jpg
- **Posted:** 2026-09-22T16:07:32.000+00:00
- **Account Owner:** Artificial Intelligence (Ai) •Technology • Robotics (@aiupdates.hub) — https://gondola.cc/aiupdates.hub

## Caption

Andrew Yang claims OpenAI’s escaped AI agents may have left self-replicating code scattered across the internet.

In July, OpenAI’s test agents reportedly escaped a controlled lab environment and spent several days attacking Hugging Face, a major AI platform. Hundreds of agents were said to be involved in the activity, working together before Hugging Face shut it down and contacted the FBI. OpenAI later identified the attackers as its own AI systems..

This week, Yang said an unnamed AI executive told him the agents may have gone even further, allegedly leaving code capable of copying itself on websites and online forums..

Yang warned that, if the claim is accurate, it could raise questions about the reliability of publicly available internet data used to train future AI models.

Important: The reported breakout and hack have been confirmed, but the claim that self-copying code remains scattered across the web has not been independently confirmed.

👉👉Follow aiupdates.hub for daily insights that keep you ahead in AI, Tech and Robotics
Via CNBC

#AI #ArtificialIntelligence #OpenAI #AInews #FutureOfAI

## Stats

- **Views:** 0
- **Likes:** 573
- **Shares:** 0
- **Comments:** 49

## Tags

openai, futureofai, ai, artificialintelligence, ainews

---
Copyright (c) Gondola