OpenAI seems to be silently re-enabling model training on Codex chats

OpenAI's Codex is silently re-enabling model training on user interactions, potentially allowing user data to be used for model improvement without explicit consent. Users can check their data sharing settings on chatgpt.com/settings/data-controls and also on privacy.openai.com/policies. A previous occurrence of this issue was noted in 2022. Users can turn off data sharing on privacy.openai.com/policies. The issue has been reported by multiple users and is not isolated. A link to check data sharing settings is provided in the original post. A link to a previous occurrence of the issue is also provided in the original post.