Unconfirmed: Mark Gurman said in posts on X on 19 November that Apple is developing an autonomous AI agent that can control devices, browse the web, and write code; Apple has not confirmed it.

In one post on X1, Gurman wrote: "NEW: Apple is developing an autonomous AI agent with computer-use capabilities, including device control, app interactions, web browsing and coding. The effort goes well beyond Siri and aims to compete with Meta’s Muse, ChatGPT’s Dots and Claude Code." That post includes a t.co link, but the destination and any underlying document were not part of the material reviewed here.

In a second post on X2, Gurman wrote: "Apple’s own job listing spells out the ambition: An AI agent that can use a computer like a person, moving between browsers, apps, terminals and services, writing and debugging code, and deciding which tools to use to complete a task. This is far beyond today’s Siri."

In a third post on X3, he added: "A particularly notable detail: Apple specifically references “general-purpose computer-use agents.” That points to AI that can autonomously navigate computer interfaces, not just invoke predefined app functions. It’s the same technology powering computer use in ChatGPT and Muse."

What the evidence establishes is limited. It shows that the @markgurman account, which has about 491,450 followers in the reviewed material, made those claims on 19 November. It does not independently establish that Apple is developing the system Gurman describes, and it does not include the text or URL of the alleged Apple job listing.

What would confirm the rumor is straightforward: an Apple announcement, a public Apple job posting containing the language Gurman described, or other original Apple documentation describing a "general-purpose computer-use" agent. What would seriously weaken it would be Apple denying the work, the cited listing proving unavailable or materially different, or later reporting showing Gurman’s summary does not match the underlying document. Apple has not commented in the material reviewed.

There is no earlier Automatica reporting in the current evidence set that specifically concerns Apple’s agent plans.

Cited posts