据彭博社消息,苹果公司正着手开发一款能够模拟人类操作计算机的专用人工智能智能体 1。该项目的核心能力涵盖浏览网页、跨应用程序协同工作、执行系统命令以及编写和调试代码 12。为构建这一旨在提供私密且符合苹果标准体验的系统,公司目前正在积极扩充工程团队,并发布了招聘高级工程经理等职位的需求 12。尽管相关报道指出该项目尚处于早期阶段,但苹果计划争取在明年年底前推出这项服务 1。
在推进自研智能体的同时,苹果也在加强对macOS系统中第三方智能体的安全管控。据悉,苹果计划在未来的macOS版本(具体提及macOS 27.3)中部署警告机制,以阻止用户授予第三方智能体完全访问权限 1。早在2024年10月,苹果已向开发者发出通知,表示将引入额外的控制措施,要求必须经过明确的用户操作才能向应用授予此类特殊级别的访问权 1。此外,受苹果意图发展自有AI代理的消息影响,Anthropic公司的股价出现了下跌 2。
Apple is reportedly developing an artificial intelligence agent capable of operating a computer like a human user. Bloomberg reports that the company has posted a job listing for a senior engineering manager to build this "computer-use AI agent." 1 The software would be able to browse the web, work across applications, execute commands, write and debug code, and invoke services. 12
The project aims to provide a private user experience consistent with Apple’s standards. 1 Apple is actively hiring engineers to form a dedicated team for this privacy-first system. 2 Although the initiative remains in its early stages, Bloomberg notes that Apple intends to launch the service by the end of next year. 1
Concurrently, Apple plans to restrict third-party agents from gaining full control over devices to enhance security. 1 Future versions of macOS, specifically mentioned as macOS 27.3, will include warnings preventing users from granting complete access permissions to external AI tools. 1 In October 2024, Apple notified developers that additional controls would require explicit user actions before applications could receive such elevated access levels. 1 Market reactions have included a decline in Anthropic’s stock price, reportedly driven by concerns over Apple’s entry into the AI agent space. 2
评论
还没有评论,欢迎留下第一条。