OpenAI has released an Apple Messages plug‑in for ChatGPT that connects a user’s iMessage inbox to the model, allowing the chatbot to read, sort, analyze, edit, draft, send, and delete messages as well as search through conversation history. The integration also works with Codex and ChatGPT Work, enabling similar functionality in professional workflows.
The plug‑in executes locally on the user’s machine; OpenAI states that it “doesn’t create an index of all someone’s messages,” although the precise technical meaning of that claim has not been elaborated. Users are advised to monitor ChatGPT’s actions and to avoid turning on persistent auto‑approval, which would eliminate the final human review before a message is sent.
- Local execution on device
- No persistent message index (per OpenAI statement)
- Compatible with Codex and ChatGPT Work
- Capabilities: sort, analyze, edit, draft, send, delete, search
- Privacy guidance: retain manual oversight, avoid persistent approval
Why this matters
The local‑only execution model suggests an attempt to limit server‑side exposure of personal message data, reducing the risk of centralized data collection. However, because the plug‑in still processes message content on the device and can act on behalf of the user, any vulnerabilities in the local implementation or misuse of auto‑approval could lead to unintended message disclosure or sending. The lack of detailed technical disclosures about how the “no index” guarantee is achieved leaves room for uncertainty, meaning users must weigh the convenience of automated messaging against the need for continued vigilance over privacy and security.
Share this article
Found this insightful? Share it with your community on Reddit, X, or copy the link.