AI Reality Check: Unexpected Risk: Chatbots with Memory

OpenAI has recently introduced a new feature: Your ChatGPT account will remember what you have done in the past and use that information to give you better prompting results. Google does something similar and includes your search results. 

While the improved quality this feature will provide is good news, what is potentially not so good is when team members use their private chatbot accounts for work. Sensitive and/or contractually restricted information, like contract documents, are uploaded and saved in memory. While this isn’t a huge risk if the person stays, it opens unexpected risks if they leave or even if a consultant or temporary worker, like an intern, uses their chatbots for company work.