Tech

Apple developing in-house AI agent capable of controlling computers

A recent job opening revealed Apple is building an artificial intelligence agent designed to operate computers, write code and execute multi-step tasks across apps.

Apple Computer Headquarters, street sign
en:User:Nishant12 / Wikimedia Commons, Public domain

Apple appears to be working on its own in-house artificial intelligence agent capable of operating a computer on a user’s behalf, according to details from a corporate job posting first reported by Bloomberg.

The job listing, posted for a senior engineering manager position, reveals that the iPhone maker is building software designed to navigate web browsers, write and debug code, run command lines, and complete multi-step tasks across multiple applications. The move indicates that Apple is seeking to keep pace with rival tech companies by entering the rapidly growing field of autonomous AI software.

Inside Apple’s AI Agent Ambitions

According to the employment listing, Apple intends for the software to function similarly to a skilled human user while maintaining strict privacy standards. “Imagine an AI agent that can use a computer the way a capable person does,” the posting states, according to PYMNTS.com. “It browses the web, works across apps, runs commands, writes and debugs code, and calls services, and it chooses the right tool for each step. It does all this privately and with the polish people expect from Apple.”

The job description further outlines plans for agents that can move “fluidly between tools” and manage complex multi-step workflows across command lines, application programming interfaces, and web browsers, TechCrunch reported. The incoming engineering leader will oversee the technical direction and delivery of the company’s agent platform, collaborating directly with internal teams in Siri, User Interface Frameworks, and Apple Foundation Models, according to details cited by The Verge.

While Bloomberg reported that Apple aims to introduce the service as early as late next year, development remains in its early stages.

Rising Competition and Security Concerns

The revelation comes as major technology companies rush to launch digital agents that interact directly with desktop operating systems. Meta introduced its personal AI agent, Muse, on Sept. 8, while OpenAI launched its “dots” AI agents on Sept. 28, according to PYMNTS.com. Anthropic released its Claude Duet application on Oct. 21, and SpaceXAI launched Grok Bot in mid-August. Additionally, Google announced that its Gemini application for Mac can now complete multistep tasks across local desktop software and web tools.

As Apple develops its proprietary agent, the company has simultaneously taken steps to restrict third-party software from taking unrestricted control of its Mac computers. Engadget reported that Apple recently cautioned developers regarding security risks associated with granting full disk access to agentic AI tools. The upcoming macOS 27.3 update will introduce warnings urging users to exercise caution before providing third-party AI tools total access to their operating system, according to Bloomberg.

Future Integration Across Devices

Apple has already begun expanding the capabilities of its virtual assistant, Siri. The company began rolling out an AI-upgraded Siri in September, enabling users to query on-screen content and execute actions across applications, according to The Verge. Furthermore, Apple is offering a Developer Preview with its upcoming iOS 27.3 update that will allow developers to integrate Siri AI features into all applications.

It remains uncertain when Apple will officially unveil its autonomous agent platform or how it will balance high-level computer automation with its privacy requirements.