On January 25, OpenAI officially tested a major upgrade to the Temporary Chat feature. This update addressed the long-standing issue of this mode being "too clean," allowing users to access personalized settings without recording conversation history, achieving a balance between privacy protection and usage efficiency.
Temporary Mode Is No Longer a "Stranger"
In previous versions, enabling "Temporary Chat" meant entering a completely blank mode: the model could not access user memory (Memory) or custom instructions (Custom Instructions). After the latest upgrade, the Temporary Chat Feature will be able to recognize and retain user preferences, including specific language styles, tones, and saved background knowledge.
Nevertheless, the core security mechanism remains unchanged: conversation records will not appear in the history list nor be used to improve OpenAI's models. This means users can use existing personalized configurations to handle sensitive tasks without worrying about these specific conversations affecting their account's long-term memory.
Along with the Upgrade: AI Age Prediction Model Launched
In addition to the feature upgrade, OpenAI recently launched the Age Prediction Model, aimed at identifying underage users through conversation behavior and limiting their exposure to sensitive topics such as violence. Although this model carries the risk of misjudgment, adult users can unlock the full experience through age verification appeals. Currently, users can simply click the "Temporary" button in the top right corner of the ChatGPT interface to access the upgraded private conversation experience.
