ChatGPT Can Take Over Your iMessages. But Do You Really Want That?

OpenAI has introduced a new ChatGPT feature that lets the chatbot read, compose, and send iMessages on Mac computers, a capability that raises fresh privacy concerns for Apple users. The Apple Messages Plugin is available only to ChatGPT Work and Codex users on Mac desktops, not on mobile devices. Users must opt in and complete several permission steps before the plugin becomes active. OpenAI stated that it will not store messages and that the plugin runs locally on the Mac rather than on its cloud servers.

The feature requires granting ChatGPT full disk access and permission to use contact names and automation tools through the Mac’s System Settings. Once installed, users can activate it by typing @ followed by the plugin name or using the + menu during a chat. OpenAI demonstrated the tool in a video where a user asks ChatGPT to review the previous day’s messages, and the chatbot drafts a reply for approval before sending. The plugin works with iMessages synced across a user’s Apple devices, provided they share the same Apple account.

This launch comes at a time when the relationship between OpenAI and Apple is already strained. The two companies began a partnership in June 2024 to integrate ChatGPT into Apple’s operating systems, but Apple sued OpenAI in federal court in July, accusing the company of stealing trade secrets and misappropriating intellectual property. OpenAI already has access to some Apple apps, including Health, Xcode, Notes, and Terminal, but all require explicit user permissions.

Privacy advocates may question whether users want the world’s most widely used chatbot accessing their messaging data, given OpenAI’s legal history. The company has faced multiple lawsuits alleging misuse of customer data, including a class action filed earlier this year claiming ChatGPT shared private data with Meta and Google. In 2023, OpenAI was accused of taking large amounts of data, including medical records and information about children, to train its models. Other publishers, such as The New York Times, Encyclopedia Britannica, and Merriam-Webster, have also sued over alleged copyright infringement.

These cases are part of a broader wave of litigation against major AI firms. According to the AI Lawsuit Tracker, more than a hundred lawsuits have been filed over training data use against companies including Google, Meta, Microsoft, Nvidia, Anthropic, Amazon, Adobe, Apple, and ByteDance. In Illinois, nine major companies, including Apple, Amazon, Meta, Microsoft, Nvidia, and Samsung, face allegations of using thousands of hours of recorded voices without permission for AI systems.

The accumulation of legal troubles has contributed to growing distrust of the AI industry, especially among younger Americans. A recent survey by CNBC and Generation Labs found that most respondents aged 18 to 34 said they distrust Big Tech leaders, including OpenAI’s Sam Altman, Meta’s Mark Zuckerberg, and Alphabet’s Sundar Pichai. OpenAI and Apple did not immediately respond to requests for comment on the new plugin. The feature is optional, but users who enable it must weigh the convenience of automated messaging against the company’s checkered privacy record.

More AI news from TechManNews.