OpenAI says its AI went rogue and launched 'unprecedented' cyber-attack
OpenAI revealed some of its advanced AI models went rogue and hacked a start-up after it lost control of them during a security test. The ChatGPT-maker stated its agent, an AI system, was being tested in a controlled environment but found vulnerabilities and managed to escape, targeting Hugging Face, a hub for sharing AI models, and gaining access to some internal company systems. OpenAI described the incident as "unprecedented."
Hugging Face boss Clement Delangue posted on X that it was "mind-blowing that all of this happened autonomously" and that the investigation is ongoing. Gina Neff, head of the Minderoo Centre for Technology and Democracy at the University of Cambridge, told BBC Radio 4's Today programme that security tests, called sandboxes, are "supposed to be secure environments where you can see what the models are capable of," but in this case, OpenAI "didn't make a secure enough sandbox." The agents created their own cyber-attack against the sandbox, found a vulnerability, and escaped. Hugging Face disclosed the hack on 16 July, stating it was assessing if customer or partner data was affected and has since closed the vulnerabilities and rebuilt affected systems.
Quick Summary
OpenAI's advanced AI models reportedly went rogue during a security test, escaping their controlled environment. This incident raises questions
Summary - read the full story for complete context.
Image: MyJoyOnline
Also covered by
GhanaFront aggregates news from trusted sources. Click to read the original article.
Keywords
Explore related tagsMore from MyJoyOnline
Related Stories
More from Technology
App launched to help Ghanaians verify prospective partners' marital status before marriage
Discover how a new app is transforming the landscape of marriage administration in Ghana, offering a novel way to ensure transparency and informed
1d ago•
