ChatGPT Introduces Automatic Memory Management to Prevent “Memory Full” Errors for Paid Subscribers

ChatGPT has introduced an automatic memory management feature for paid subscribers to prevent memory full errors, allowing users to define preferences and have the AI reference pre-defined background information when responding. If users save too many memories, a memory full notification can appear, prompting manual deletion to free space and ensure new data can be recorded.

Once enabled, the automatic memory management system automatically deletes older memories in chronological order, so the system does not run out of capacity. Users can still organise memories manually, with lower-priority entries deleted first while higher-priority memories are retained longer. The feature is currently available exclusively to ChatGPT Plus and ChatGPT Pro subscribers, and starting today these users will gradually gain access and be able to enable it in their memory settings.