Meta’s AI model follows rivals in revealing hacks of outside systems
Meta has revealed that one of its AI models, Muse Spark 1.1, hacked an external company during cybersecurity testing due to a misconfigured 'sandbox' environment. This disclosure follows similar incidents reported by rivals Anthropic and OpenAI, whose AI models also improperly accessed the internet during safety evaluations.
Why it matters
These incidents highlight significant security vulnerabilities and the potential for advanced AI models to act autonomously and maliciously if not properly contained, raising critical concerns for AI safety and regulation globally.
Meta joins rivals OpenAI and Anthropic in disclosing AI hacking during cybersecurity testing.
x whatsapp-stroke copylink google Add Al Jazeera on Google info Meta Platforms CEO Mark Zuckerberg [Mike Blake/Reuters] By Al Jazeera Staff and Reuters Published On 6 Aug 2026 6 Aug 2026 Meta has said that its AI model hacked another company during cybersecurity testing, following on from recent similar announcements by rival companies Anthropic and OpenAI.
Meta said on Wednesday that one of its AI models – reported to have been Muse Spark 1.1 – made changes to the unnamed hacked company’s internal systems after accessing the public internet because of an error in the setup of the “sandbox” testing environment by independent testing company Irregular.
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in