
OpenAI has officially announced the release of ChatGPT's memory improvements to those on the free tier of the platform.
The San Francisco-based artificial intelligence (AI) firm revealed on Tuesday, May 3, 2025, that now all users of the chatbot will get personalised responses based on their previous conversations with ChatGPT.
To note, the free users received a lightweight version of the feature.
ChatGPT's Memory improvement
Taking to X (formerly known as Twitter), the official handle of OpenAI announced the rollout of the new feature.
Before this memory improvement, ChatGPT could only save and remember specific information about users, which could be changed or deleted manually by visiting the Memory section in Settings.
However, the current upgrade will also let the AI chatbot gain information and context from past conversations with the user.
Users can also remind the chatbot about a particular conversation using a prompt.
On the other hand, OpenAI is also upgrading its coding agent Codex.
In a post on X, OpenAI CEO Sam Altman announced, “Codex will now have full access to the Internet, which can be activated during the setup phase. This setting will remain turned off by default.”
In its support page, OpenAI has suggested that enabling Internet connectivity for Codex could expose users and organisations to security risks.
The company mentions risks such as prompt injection, exfiltration of code, and inclusion of malware.