AI Trends

Apple Reportedly Developing Its Own AI Agent Capable of Controlling Devices, Browsing the Web, and Writing Code

Sina TechnologyUpdated Nov 20 07:49

AI summary

Apple is developing an autonomous AI agent designed to control devices and applications, browse the web, write code, and perform multistep tasks across applications, browsers, command lines, and APIs on users' behalf. Job postings indicate that the system will also coordinate multiple sub-agents, with an emphasis on permission management, sandbox isolation, and operational transparency. The project remains at an early stage, and a potential launch as early as late 2027 has not yet been confirmed.

Why read this

AI score 53

Job postings reveal the capabilities and safety design of Apple's autonomous agent, offering insight into its product vision for executing tasks across applications.

Topics

Tags

Body · AI translation

Add public testing promotional copy Add public testing promotional copyend Add app-exclusive traffic redirection

IT Home, November 20 — Bloomberg’s Mark Gurman reported today that Apple is developing autonomous AI agents designed to perform tasks on users’ behalf, including controlling devices, operating apps, browsing the web, and programming software, further entering the AI agent market contested by companies such as OpenAI, Anthropic, and Meta.

Gurman found that Apple has posted job openings on its official website, including one for a senior engineering manager titled “Agentic AI - Proactive.” The listing says Apple is assembling a team to develop AI assistants capable of carrying out tasks on users’ behalf. The systems will be able to write code, browse the web, control applications, and operate devices.

Apple described in the job listing how its future AI agents would work: users could have the AI operate a computer like a skilled human, switching among browsers, documents, terminals, code editors, and various services while independently selecting the appropriate tools to complete tasks.

The system would not only execute individual instructions but also plan and complete multi-step tasks spanning applications, browsers, command lines, and APIs, while coordinating multiple sub-agents specializing in different jobs. Apple also placed particular emphasis on privacy protection, saying that agents capable of running code, using user credentials, and taking action on users’ behalf require strict permission management, sandbox isolation, and operational transparency.

However, the project remains at an early stage of development. Gurman said Apple had considered launching the service as early as late 2027, but that timeline has not yet been finalized. An Apple spokesperson did not respond to Bloomberg’s request for comment.

Gurman said Apple’s new CEO, John Ternus, has been pushing software engineering chief Craig Federighi to accelerate AI development and adopt a more aggressive strategy. Autonomous AI agents are seen as an important part of Apple’s efforts to advance that strategy.

Several competing products have already emerged in the AI agent market. Meta’s Muse recently attracted large numbers of users with capabilities such as booking restaurants, arranging travel, and generating news briefings; Anthropic is expanding user testing of its own agent, Duet, while OpenAI continues to broaden the functionality of Dots. Google adjusted its agent this week, renaming it Tasks and adding capabilities such as making phone calls on users’ behalf.

Meanwhile, OpenAI, Anthropic, and Elon Musk’s SpaceXSI have also introduced developer programming tools such as Codex, Claude Code, and Cursor, respectively. Gurman noted that AI agents have already secured a prominent position in Apple’s App Store. As of Thursday local time, Meta’s Muse had long ranked second on the free-app chart, behind only TikTok, while ChatGPT and Claude were also among the top five free apps.

While advancing development of its own agent, Apple is also adjusting macOS permission management for third-party applications. The forthcoming macOS 27.3 beta adds permission warnings for third-party AI agents, requiring users to reconfirm when granting such applications permission to control their Mac. The change increases the number of steps required for ChatGPT users to enable third-party computer-control tools from four to 14. Some AI companies and users believe that, beyond protecting system security, this may also make it harder for competitors to integrate with macOS.

Apple, however, has explained the changes as measures to strengthen the protection of user data. On November 16, Apple announced planned permission changes for macOS 27.3 Beta 2 involving device control, data access, and full disk access, with an emphasis on improving the transparency of permission management and providing more granular access-control options.

In addition to a standalone AI agent, Apple is also advancing Siri’s integration with third-party applications. The forthcoming iOS 27.3 will give developers deeper Siri AI integration capabilities, allowing the voice assistant to invoke external applications to complete tasks such as booking a car or a restaurant.

At this stage, however, Siri’s integration with third-party applications still differs from a fully autonomous AI agent. The latter can independently plan and execute complex tasks in the background across multiple applications and services, reducing the need for users to repeatedly provide instructions during execution. article_adlist[

Related reading: ]article_adlist Statement: This is an exclusive Sina.com article. Reproduction without authorization is prohibited. tech_web_article_v2015_block_bottom Non-targeted 300*250 button 17/09 wenjing begin

Non-targeted 300*250 button end

Source

Sina Technology