ChatGPT can now send texts for you with new Apple Messages plugin
OpenAI has released an Apple Messages plugin for ChatGPT, letting users link their Messages inbox directly to the chatbot. The move extends ChatGPT's reach into personal communications, enabling the AI to sort, analyse, edit, draft and even send texts on a user's behalf, while also raising fresh privacy questions given how much sensitive conversation data would flow through the system.
The plugin works with ChatGPT's Codex and ChatGPT Work products as well as personal accounts, meaning it can be used professionally too. OpenAI says it can suggest follow-up replies based on previous messages, delete texts, or search through message history for specific information. The company told Bloomberg the plugin runs locally and does not create a full index of a user's messages, though it has not detailed exactly what that entails. OpenAI is also advising against enabling "persistent approval" for message sending, warning that doing so removes a user's final chance to review a message before ChatGPT sends it in their name.
- OpenAI launches ChatGPT plugin to access and send Apple Messages texts
- Plugin can sort, edit, draft, delete and send messages for users
- OpenAI warns against auto-approving sends; privacy details remain unclear
New here? Start with this
OpenAI's ChatGPT can now be linked to Apple's Messages app, the built-in text messaging system on iPhones and Macs. The new plugin lets the AI read a user's texts and, with permission, organise them, write replies, and send messages on the user's behalf, rather than ChatGPT being limited to a separate chat window as before.
OpenAI is the company behind ChatGPT, one of the best-known AI chatbots, and it has been steadily building tools that connect the assistant to everyday apps and tasks. Apple's Messages app is where many iPhone and Mac users hold personal and work conversations, so linking the two means an AI system would have far deeper access to private exchanges than typical chatbot use.
This matters because it marks a shift from AI as something people actively query towards AI that can act on their behalf inside sensitive, personal spaces. It raises questions about how much conversation data is handled, stored or reviewed, and about the safeguards in place to stop messages being sent without a person's final say-so.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Users gain significant productivity benefits from having AI handle routine messaging tasks like drafting, sorting and analysing texts. The plugin maintains user control through required message review before sending, ensuring humans retain final authority over communications sent in their name. For busy professionals, delegating these administrative elements of messaging could free considerable time, and the feature remains optional with local processing suggesting privacy safeguards have been considered.
The case against
Personal messages contain sensitive information—financial details, health concerns, relationship matters—now flowing through additional systems, raising legitimate privacy and security concerns. Even with assurances about local processing, users may not fully understand what data is accessed or how breaches could expose intimate conversations. Beyond data security, normalising AI sending messages in one's name risks undermining authenticity in human relationships and creates potential for real harm if the system errs or is misused.