FATA News Alert
Contributing Author(s): Emily Chee, CPA, CMA, CITP, CISA, CFF, CFE; Tony Harris CISSP, SC-100, CRISC
In the News: Hacker Injects False Memories in ChatGPT to Steal User Data Indefinitely¹
Background:
In 2024, a vulnerability was discovered related to OpenAI’s long-term conversation memory functionality. OpenAI’s long-term conversation memory was a feature introduced in February 2024 and became more broadly available in September 2024. This feature retains user information to enhance future conversations. Details such as a user’s age,