News

Apple reportedly developing proactive AI agent for coding and autonomous computer use

Published: | Posted By:

The era of the passive voice assistant is ending as Apple pivots toward a proactive AI agent capable of autonomous computer use. This shift represents a move toward "agentic AI," where the system does not merely answer questions but actively executes complex, multi-step tasks across various applications and operating systems.

This ambition is evidenced by recent job listings for a senior engineering manager tasked with leading the development of a proactive, agentic AI assistant. According to these listings, Apple envisions an agent that can use a computer the way a capable person does, browsing the web, running commands, and writing or debugging code. The goal is to create a general-purpose computer-use agent that can coordinate specialized sub-agents to plan and carry out tasks across browsers, command lines, and APIs, all while maintaining the polish and privacy expected of the brand.

The Competitive Landscape

While Apple is still in the early stages of development, the market for proactive agents has already exploded. Meta’s Muse assistant has seen significant success, recently occupying the number two spot among free apps on the App Store. Other frontier labs have launched similar tools, including OpenAI’s Dots, Anthropic’s Claude Duet, and Google’s Tasks—the latter of which can even make phone calls on a user's behalf. The pressure to compete is apparent not only in the App Store rankings but also in the market, as shares of Meta and Anthropic saw slight declines following reports of Apple's internal efforts.

This push for autonomy comes amid reported internal pressure within Apple's leadership. New CEO John Ternus has reportedly been pushing software engineering chief Craig Federighi to adopt a bolder approach to AI and accelerate the delivery of these capabilities. While Bloomberg suggests Apple may be aiming for a launch as soon as the end of 2027, the current state of Siri AI remains a distant contender compared to the proactive nature of the competition.

Security and Ecosystem Control

Because an agent capable of running code and using credentials could pose significant risks, Apple is emphasizing that safety, privacy, and security must be central to the design. The job listings specifically mention the need for strong permission models and transparency, as well as the use of sandboxing—a process that allows software to be tested in isolation to prevent it from breaking into the broader operating system or external networks.

Interestingly, as Apple builds its own agent, it appears to be making it more difficult for third-party agents to operate on its hardware. In the beta version of macOS 27.3, Apple has introduced new warning screens for users granting third-party agents access to their Macs. Furthermore, the process for enabling ChatGPT’s computer-use tools has been increased from four steps to 14. While Apple positions these changes as necessary security measures, some observers view them as a strategic attempt to slow down competitors.

These developments suggest a future where the operating system is no longer just a platform for apps, but a managed environment for an autonomous agent. While the upcoming iOS 27.3 release lays the groundwork for Siri AI to collaborate more deeply with third-party apps for tasks like booking reservations, a full-fledged agent would represent a fundamental change in how users interact with their devices. The success of this transition will likely depend on whether Apple can close the existing technology gap while maintaining the strict privacy standards that define its ecosystem.

Read more from Ninjalane
Hardware Reviews, News and More...