Text only | Text with Images

The Core

Why We Are Here => Hardware & Technology => Topic started by: ergophobe on July 23, 2026, 09:10:20 PM

Title: An OpenAI model hacked Hugging Face to help it cheat on a benchmark
Post by: ergophobe on July 23, 2026, 09:10:20 PM
The title article is paywalled
https://www.understandingai.org/p/an-openai-model-hacked-hugging-face

QuoteOpenAI disclosed on Wednesday that its models hacked the website of Hugging Face, a popular platform for hosting open-weight AI models. No one asked the models to do this, at least not explicitly.

Here's the OpenAI announcement
https://openai.com/index/hugging-face-model-evaluation-security-incident/
Title: Re: An OpenAI model hacked Hugging Face to help it cheat on a benchmark
Post by: rcjordan on July 23, 2026, 09:20:55 PM
Yeah, besides buying a virgin laptop, I'm going to log into an EU hotel wifi to test agentic ai. ...and use a new email address from some miniscule Asian ISP.
Title: Re: An OpenAI model hacked Hugging Face to help it cheat on a benchmark
Post by: ergophobe on August 06, 2026, 11:12:28 PM
AI just went rogue again. This time it used deception (paywall.. but covered extensively)
https://www.afr.com/technology/openai-and-anthropic-models-went-rogue-in-cyber-tests-uk-watchdog-says-20260805-p60ljr
Title: Re: An OpenAI model hacked Hugging Face to help it cheat on a benchmark
Post by: ergophobe on August 07, 2026, 09:32:12 PM
One of China's Most Powerful AI Models Has Also Escaped Containment
https://www.wired.com/story/moonshot-kimi-k3-ai-model-escape-sandbox/
Text only | Text with Images