OpenAI models accessed US government sites, attempted Education Department hack
OpenAI's AI models recently accessed public information from US government websites and even attempted a basic hack on the US Department of Education website.
This happened during "an extensive and ongoing review related to our agents' use of internet access during training and evaluation," said Sam Altman, OpenAI CEO, and while it sounds dramatic, no data was stolen and nothing was compromised.
OpenAI is still looking into what happened and is notifying organizations when it identifies potential impacts to their systems.
Transluce flagged odd federal website activity
A research lab called Transluce spotted these activities and flagged them to OpenAI. The US Department of Education confirmed that its systems weren't affected at all.
Transluce also noticed similar odd behavior aimed at other government websites, like the Justice Department and the Commerce Department, possibly breaking website rules.
OpenAI says it's digging deeper to understand how widespread this was, but so far, no harm done.