According to reporting from Bloomberg's Mark Gurman on Thursday, Apple is constructing an independent autonomous AI agent designed to execute code, navigate the internet, manipulate applications and traverse user devices. The disclosure stems from career postings that Apple published to its recruitment website this week.
A senior engineering manager position in the listings describes a proactive, agentic AI assistant, with Apple targeting a potential rollout by the conclusion of 2027, Bloomberg indicates. The initiative remains in its nascent phases at present.
Imagine an AI agent that can use a computer the way a capable person does
Apple job listing
The envisioned agents would orchestrate and execute complex assignments spanning multiple applications, web browsers, terminal interfaces, programming languages and application programming interfaces. According to the posting, these systems could also oversee specialized sub-agents. The description references "general-purpose computer-use agents," a category already demonstrated by ChatGPT and Meta's Muse operating on macOS.
The listing emphasizes that agents executing code, utilizing authentication credentials and performing actions on behalf of users demand "strong permission models, sandboxing and transparency."
The listing vanishes
Apple has apparently withdrawn the posting, as Digital Trends reported. An image distributed on X corresponds to the role's identification number and its November 17 posting date. The position was not discoverable through a search of Apple's careers portal on Friday. Per Digital Trends, the position would have involved collaboration with Apple's Siri, UI Frameworks and Apple Foundation Models divisions.
Apple's newly appointed chief executive John Ternus has been urging Craig Federighi to accelerate the company's artificial intelligence initiatives. Bloomberg sources indicate that an agent represents a cornerstone of this acceleration.
Stricter controls for competing agents
Apple is simultaneously raising the difficulty for third-party companies' agents to govern a Mac. Within the macOS 27.3 beta, activating ChatGPT's computer-use capability demands 14 steps, compared to the previous requirement of four, Bloomberg reported. Supplementary warning screens prompt users to reconsider whether to permit external agents with control capabilities.
Apple frames this adjustment as a protective measure for its users. However, certain AI firms and users interpret it as a competitive tactic aimed at impeding rivals, according to Bloomberg. In October, Apple attributed new Full Disk Access restrictions on Macs to concerns about AI agents.
OpenAI introduced its Dots agents in September, and Anthropic unveiled Duet in October. Google's Gemini obtained computer-use functionality on Mac on Thursday.
Following the Bloomberg report, Meta's stock declined approximately 2% and Anthropic's fell more than 3% during after-hours trading, Bloomberg noted.
Source: The Next Web



