Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

Bloomberg: Apple Is Building an AI Agent for Device Control, Apps, Browsing and Coding

Bloomberg reports that Apple is building an AI agent capable of controlling devices, operating apps, browsing the web and writing code, with follow-up coverage framing it as a proactive rival to OpenAI and Anthropic. The reports do not name the product, set a timeline, or specify which devices and apps it would support.

Published
彭博社:苹果正研发可控制设备、操作应用并编写代码的AI智能体
Image source: apple.com

Bloomberg reported on November 19 that Apple is working on an AI agent capable of controlling devices, operating apps, browsing the web and writing code. The report was quickly picked up by other outlets, with Crypto Briefing framing the effort as a "proactive" agent built to compete with OpenAI and Anthropic.

The capability list in the reporting is specific even though the sourcing is thin: device control, app operation, web browsing and coding. Taken together, those four areas describe something more than a question-and-answer assistant — they point to a system that acts on a user's behalf across applications rather than just responding inside a chat window.

The word "proactive" is the most consequential detail in the coverage. It suggests an agent that does not wait to be asked but decides on its own when to step in, although that reading comes from the outlets' framing rather than from any confirmed Apple description of the product.

For Apple, pushing agent permissions down to the device and app layer is a directional bet. A chat assistant can be swapped out; an agent that holds device control and app operation becomes the entry point to how users get things done, which is the same layer OpenAI and Anthropic have been moving toward.

The reporting is also conspicuously incomplete. It does not name the product, give a release schedule, or specify which system versions, devices or apps would be supported, and there is no word on pricing or regional availability.

That makes this closer to a directional disclosure than a product launch. It shows Apple's agent work has advanced to the point of a concrete capability list, but the distance between "in development" and something users can actually run is still filled with productization and privacy-permission questions.

Three things are worth watching next: whether Apple demonstrates agent features in an upcoming system update, how device-control permissions are opened up inside its privacy framework, and whether the capability is folded into the existing voice assistant rather than shipped as a separate product.

Why it matters

If the reporting holds, Apple's competitive focus shifts from model capability to system-level task execution. That would squeeze standalone AI assistants on its devices.

AppleAgent
Back to realtime news

Nearby Updates

All