How ChatGPT’s memory stores and uses your personal data

ChatGPT, developed by OpenAI, stores user preferences and conversation details to personalize future responses. This feature, known as memory, operates on two levels: saved memories, which include information explicitly requested or identified as useful by the system, and conversation history, which allows the assistant to retrieve themes and data from previous interactions even without a specific saved item.

Since June 2026, OpenAI has begun presenting this information in a categorized summary, replacing the previously displayed simple list. Users can view stored information by navigating to Settings > Personalization > Memory. Depending on the account type, this screen displays either a general summary or a detailed list of saved items.

How to edit or remove information

If you want to know what ChatGPT remembers about you, simply ask, “What do you know about me?” The assistant can provide either a summary or a list of memories.

• To correct data, type the updated information directly into the conversation, select a section of the summary and replace it, or check the “do not mention this again” option when available.
• To delete all memories at once, use the “delete and disable memory” option in the three-dot menu on the memory screen. This removes the summary and disables the feature, but the original conversations that generated the memories remain intact. For complete deletion, you must also delete the corresponding conversation history.

The difference between memory and model training

The memory setting controls only what ChatGPT remembers about the user. The training option, found in Settings > Data Control, determines whether OpenAI can use conversation content to improve its models. Disabling one does not affect the other.

When ChatGPT uses personal information in its responses, an icon may indicate the source, such as a previous conversation, a saved memory, a custom instruction, or an uploaded file. OpenAI notes that these indicators do not necessarily show all factors that influenced the response.

Frequently Asked Questions

How does ChatGPT’s memory work?

According to OpenAI’s Help Center, memory combines saved data with information retrieved from conversation history to personalize future responses.

How can I see what ChatGPT knows about me?

Go to Settings > Personalization > Memory or ask the assistant directly. The result may be a categorized summary or a list of saved items.

How do I correct information saved by ChatGPT?

Edit the text in the conversation, replace a section of the summary, or use the “do not mention this again” option when available.

How do I delete all ChatGPT memories?

Select “delete and disable memory” in the three-dot menu. To completely remove the data, you must also delete the conversations that generated the memories.

Does deleting a conversation also delete ChatGPT’s memory?

No. Deleting a conversation does not automatically remove the memories created from it; both actions must be performed separately.

Are ChatGPT’s memory and training the same setting?

No. Memory defines what the assistant remembers about the user, while training determines whether OpenAI can use conversations to improve its models.