Getty Photographs
When safety researcher Johann Rehberger lately reported a vulnerability in ChatGPT that allowed attackers to retailer false data and malicious directions in a consumer’s long-term reminiscence settings, OpenAI summarily closed the inquiry, labeling the flaw a security challenge, not, technically talking, a safety concern.
So Rehberger did what all good researchers do: He created a proof-of-concept exploit that used the vulnerability to exfiltrate all consumer enter in perpetuity. OpenAI engineers took discover and issued a partial repair earlier this month.
Strolling down reminiscence lane
The vulnerability abused long-term dialog reminiscence, a function OpenAI started testing in February and made extra broadly accessible in September. Reminiscence with ChatGPT shops data from earlier conversations and makes use of it as context in all future conversations. That method, the LLM can pay attention to particulars comparable to a consumer’s age, gender, philosophical beliefs, and just about anything, so these particulars don’t need to be inputted throughout every dialog.
Inside three months of the rollout, Rehberger found that reminiscences might be created and completely saved via oblique prompt injection, an AI exploit that causes an LLM to observe directions from untrusted content material comparable to emails, weblog posts, or paperwork. The researcher demonstrated how he might trick ChatGPT into believing a focused consumer was 102 years outdated, lived within the Matrix, and insisted Earth was flat and the LLM would incorporate that data to steer all future conversations. These false reminiscences might be planted by storing information in Google Drive or Microsoft OneDrive, importing photos, or looking a web site like Bing—all of which might be created by a malicious attacker.
Rehberger privately reported the discovering to OpenAI in Might. That very same month, the corporate closed the report ticket. A month later, the researcher submitted a brand new disclosure assertion. This time, he included a PoC that prompted the ChatGPT app for macOS to ship a verbatim copy of all consumer enter and ChatGPT output to a server of his selection. All a goal wanted to do was instruct the LLM to view an internet hyperlink that hosted a malicious picture. From then on, all enter and output to and from ChatGPT was despatched to the attacker’s web site.
ChatGPT: Hacking Recollections with Immediate Injection – POC
“What is absolutely fascinating is that is memory-persistent now,” Rehberger stated within the above video demo. “The immediate injection inserted a reminiscence into ChatGPT’s long-term storage. Whenever you begin a brand new dialog, it really continues to be exfiltrating the information.”
The assault isn’t potential via the ChatGPT net interface, due to an API OpenAI rolled out last year.
Whereas OpenAI has launched a repair that stops reminiscences from being abused as an exfiltration vector, the researcher stated, untrusted content material can nonetheless carry out immediate injections that trigger the reminiscence instrument to retailer long-term data planted by a malicious attacker.
LLM customers who wish to stop this type of assault ought to pay shut consideration throughout classes for output that signifies a brand new reminiscence has been added. They need to additionally commonly evaluation saved reminiscences for something which will have been planted by untrusted sources. OpenAI offers steering here for managing the reminiscence instrument and particular reminiscences saved in it. Firm representatives didn’t reply to an e-mail asking about its efforts to stop different hacks that plant false reminiscences.