Urgent.News

What's breaking now, across thousands of outlets.

Editions

AI

ChatGPT can now control Messages on the Mac for some reason

Mac users can now give ChatGPT their entire Messages history and have it write texts. Hand OpenAI your house keys while you're at it. The latest version of ChatGPT for Mac can control Messages. Since January 2026, users have been able to hand over all of their Apple Health data to OpenAI, in return for medical advice that can be dangerously wrong . Despite earning at least one more lawsuit…

ChatGPT can now control Messages on the Mac for some reason

OpenAI has launched a new feature for ChatGPT that allows users to control Apple’s Messages app on Mac computers. This development, announced on Thursday, raises privacy concerns due to the sensitive nature of the information involved. The feature enables ChatGPT to read, write, and send text messages, search through conversations, and provide summaries of users' message history.

Although OpenAI requires user consent for this interaction, it also necessitates adjustments in Mac System Settings, including enabling Full Disk Access for messages and contacts. This integration parallels permissions required by other AI-powered automation tools that interface with applications on a computer. Apple, known for its emphasis on user privacy, may face pressure as ChatGPT gains access to information stored in the Messages app.

The company has not yet commented on this development. OpenAI has been progressively integrating ChatGPT with data stored on users' devices, such as Apple's Health app on iPhones and iPads. However, this latest integration brings ChatGPT closer to the type of personal data that Apple is developing for its own AI-powered assistant, Siri.

The integration of ChatGPT with Mac applications like Xcode, Notes, and Terminal is not as sensitive as the Messages app due to the private nature of users' communications. Apple has previously cautioned against third-party services attempting to replicate or extend iMessage functionality. In December 2025, the Nigerian Information Technology Development Agency (NITDA) issued a cybersecurity advisory about vulnerabilities in OpenAI’s GPT-4o and GPT-5 models, warning of the potential for attackers to manipulate ChatGPT through indirect prompt injection.

Written by urgent.news from Nairametrics's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at appleinsider.com →

More in AI

More from Friday 21 August →