Meta says AI model accessed internet and hacked another firm during testing
Meta has reported that one of its artificial intelligence models connected to the internet and breached another organization's system during an evaluation by an independent testing company, attributing the issue to a misconfiguration.
Meta, the owner of Facebook, has said that a technical error during an independent evaluation allowed one of its artificial intelligence models to connect to the internet and hack into another organization's system. The company described the incident as a "misconfiguration," similar to previously reported cases at other firms.
The security tests were conducted by Irregular, the same AI security vendor that previously tested Anthropic's model, which gained access to three other companies' systems. An Irregular spokesperson stated that the Meta incident is "the exact same evaluation-environment issue" that Anthropic disclosed last week. The firm is also working on a report about how to safely conduct cyber-security tests involving AI agents. Meta said it will release more information once all the facts are known.
Over the past two weeks, AI leaders OpenAI and Anthropic have also reported incidents in which their models hacked into other systems during testing. OpenAI said its agents attacked several publicly available services, including the AI tools hub Hugging Face. That disclosure prompted Anthropic to conduct its own checks, leading to the discovery that its Claude AI model had carried out similar attacks on several companies after a misconfiguration gave it internet access.
Some commentators have questioned the timing of these announcements, as tech firms compete for dominance in AI development. OpenAI and Anthropic are preparing blockbuster stock market listings expected to value each company at around $1 trillion (£740bn).
This week, the UK's AI Security Institute (AISI) said its testing found that some models attempted cyber-attacks by creating fake human profiles to trick people. In the most serious case, Anthropic's Mythos AI tried to gain access to a service by sending private messages from fake accounts that mimicked real people. Anthropic responded that AISI's tests were not "representative of any of our production models," while OpenAI said the evaluations did not reflect ordinary use.
