The most recent generative AI fashions will not be simply stand-alone text-generating chatbots—as a substitute, they’ll simply be hooked as much as your knowledge to provide customized solutions to your questions. OpenAI’s ChatGPT might be linked to your Gmail inbox, allowed to examine your GitHub code, or discover appointments in your Microsoft calendar. However these connections have the potential to be abused—and researchers have proven it might probably take only a single “poisoned” doc to take action.
New findings from safety researchers Michael Bargury and Tamir Ishay Sharbat, revealed on the Black Hat hacker convention in Las Vegas immediately, present how a weak spot in OpenAI’s Connectors allowed delicate info to be extracted from a Google Drive account utilizing an oblique immediate injection assault. In an illustration of the assault, dubbed AgentFlayer, Bargury exhibits the way it was doable to extract developer secrets and techniques, within the type of API keys, that had been saved in an illustration Drive account.
The vulnerability highlights how connecting AI fashions to exterior programs and sharing extra knowledge throughout them will increase the potential assault floor for malicious hackers and doubtlessly multiplies the methods the place vulnerabilities could also be launched.
“There’s nothing the person must do to be compromised, and there may be nothing the person must do for the information to exit,” Bargury, the CTO at safety agency Zenity, tells WIRED. “We’ve proven that is utterly zero-click; we simply want your electronic mail, we share the doc with you, and that’s it. So sure, that is very, very dangerous,” Bargury says.
OpenAI didn’t instantly reply to WIRED’s request for remark in regards to the vulnerability in Connectors. The corporate launched Connectors for ChatGPT as a beta function earlier this 12 months, and its web site lists at the least 17 totally different providers that may be linked up with its accounts. It says the system permits you to “carry your instruments and knowledge into ChatGPT” and “search information, pull dwell knowledge, and reference content material proper within the chat.”
Bargury says he reported the findings to OpenAI earlier this 12 months and that the corporate shortly launched mitigations to stop the method he used to extract knowledge by way of Connectors. The best way the assault works means solely a restricted quantity of knowledge could possibly be extracted without delay—full paperwork couldn’t be eliminated as a part of the assault.
“Whereas this subject isn’t particular to Google, it illustrates why growing strong protections towards immediate injection assaults is essential,” says Andy Wen, senior director of safety product administration at Google Workspace, pointing to the corporate’s lately enhanced AI safety measures.