OpenAI has released a new plugin for ChatGPT that allows users to connect their Apple Messages inbox to the chatbot. According to the company, the plug-in enables users to sort, analyze, and edit messages directly within ChatGPT.
Features of the plugin
- Review and analysis: ChatGPT can examine recent and past messages and suggest follow-up replies based on messages received the previous day.
- Delete and edit messages: Users can ask the chatbot to delete messages, draft messages, and send messages on their behalf.
- Search message history: The plugin can search deep into message history to surface information that might otherwise be hard to find.
- Professional use: OpenAI says the plugin works with Codex and ChatGPT Work, enabling business as well as personal use cases.
A short promotional clip published by the company shows a user asking ChatGPT to propose follow-up messages to contacts based on the prior day’s received messages.
Privacy considerations
The new capability raises privacy concerns. OpenAI told Bloomberg that the plugin runs locally on a user’s machine and that it “doesn’t create an index of all someone’s messages.” The company’s statement does not fully clarify what that entails in practice. TechCrunch has reached out to OpenAI for further details.
Message sending and approval
OpenAI advises users to monitor what ChatGPT does when it is allowed to send messages and discourages enabling persistent approval. The company warns that persistent approval “removes your final chance to review a message before ChatGPT sends it as you.”
Why it matters
The plugin could improve convenience and productivity for both personal and professional users, but because it touches users’ private message streams and relies on local processing, questions remain about data handling, storage, and whether any metadata or indexes are created. Given the limited public details so far, users and journalists are seeking more clarity from OpenAI.
Next steps
TechCrunch has contacted OpenAI for additional information about how the plugin operates and how data is protected. Meanwhile, OpenAI’s guidance is for users to manually review messages generated by the chatbot and to avoid enabling automatic send permissions without careful consideration.
