According to Bloomberg, OpenAI rolled out a feature that lets ChatGPT access Apple's Messages app on Mac. The AI can search conversations, summarize threads, and compose or send texts directly—all without leaving the ChatGPT interface.
This hits Apple's most sensitive claim: that your private messages stay private. When an outside AI can inspect and act on your iMessages, Apple's privacy promise shifts from hardware branding to permission design and user comprehension. The question isn't whether the feature works. It's whether users understand what they're allowing, and whether Apple can defend that choice when something goes wrong.
Apple will face pressure on how it handles this in macOS and its broader AI strategy through 2026. The pressure points: how clearly app permissions are explained, whether message access gets tightly scoped, and whether Apple steers users toward its own assistant for the same tasks instead.
Users will love the convenience until a chatbot is skimming old arguments, one-time passwords, or family logistics to save thirty seconds. Then Apple owns the trust problem even though OpenAI built the feature. The company has spent years positioning itself as the privacy alternative to Google and Meta. That positioning only works if users believe their most sensitive data stays off-limits to third parties. ChatGPT reading iMessages tests whether that belief survives contact with actual convenience.
Filed to the Technology desk · Aug 20