ChatGPT is getting a new way to interact with Apple Messages. OpenAI has introduced an Apple Messages plug-in that allows users to connect their Messages inbox with ChatGPT and use AI to manage conversations.
With the integration, users can ask ChatGPT to search through messages, organize conversations, draft replies and even send messages on their behalf.
What can ChatGPT do with Apple Messages?
The new integration is designed to make everyday messaging easier. Users can ask ChatGPT to find information buried in their message history, suggest follow-up messages or edit existing text.
ChatGPT can also delete messages and send new messages for users. The feature works with other OpenAI tools, including Codex and ChatGPT Work, expanding its potential beyond personal conversations.
For example, a user could ask ChatGPT to review recent conversations and suggest replies to people they have recently contacted.
Privacy is an important concern
Giving an AI assistant access to personal messages naturally raises privacy questions.
OpenAI says the Messages plug-in runs locally on the user’s device and does not create a complete index of their messages. According to the company, ChatGPT only accesses messages when a user specifically asks it to do so.
Setting up the integration requires Full Disk Access, which gives the application permission to read and write files on the device.
OpenAI also says message content and desktop conversations are stored locally by default rather than being sent to its servers. Users who choose to save conversations in the cloud are subject to the applicable data retention policies.
Users should review messages before sending
One of the biggest changes is that ChatGPT can send messages on a user’s behalf. OpenAI recommends that users keep an eye on the AI’s actions and review messages before they are sent.
The company also warns that enabling persistent approval could remove the user’s final opportunity to review a message before ChatGPT sends it.
The update is another example of AI assistants moving beyond answering questions toward taking actions for users. As these capabilities expand, convenience will need to be balanced with privacy and user control.
Read More on VitalStack
Enjoyed this article?
Subscribe for weekly deep-dives on AI and health — straight to your inbox.