ChatGPT Work Brings Desktop Automation, Memory and Governance Into the AI Workflow
ChatGPT Work Brings Desktop Automation, Memory and Governance Into the AI Workflow
ChatGPT Work 将桌面自动化、记忆功能与治理引入 AI 工作流
OpenAI is expanding ChatGPT beyond chat with ChatGPT Work, a cross-platform work environment that includes a desktop agent able to interact with local applications, files and browser content. The change matters because it moves ChatGPT closer to an operational role: not only explaining how to complete a task, but potentially clicking, typing, moving files and staying engaged with a project over time. OpenAI 正在通过 ChatGPT Work 将 ChatGPT 的功能从单纯的聊天扩展开来。这是一个跨平台的工作环境，包含一个能够与本地应用程序、文件和浏览器内容进行交互的桌面智能体。这一变化意义重大，因为它使 ChatGPT 更接近于操作角色：它不仅能解释如何完成任务，还可能直接进行点击、输入、移动文件，并长期参与项目。
In OpenAI’s official announcement on ChatGPT Work, the company describes a unified experience across web, mobile and desktop. The desktop app combines Chat, Work and Codex, while its built-in browser and local computer capabilities are intended to support more contextual, end-to-end work. OpenAI’s terminology centers on ChatGPT Work and Computer Use. “Computer History,” the name used in the originating signal, is not the feature name used in the official announcement. 在 OpenAI 关于 ChatGPT Work 的官方公告中，该公司描述了一种跨网页、移动端和桌面的统一体验。桌面应用程序结合了 Chat（聊天）、Work（工作）和 Codex（代码）功能，其内置浏览器和本地计算机能力旨在支持更具情境化的端到端工作。OpenAI 的术语主要围绕 ChatGPT Work 和 Computer Use（计算机使用）。最初信号中使用的名称“Computer History”（计算机历史）并非官方公告中使用的功能名称。
The underlying shift is significant for developers and knowledge workers. A chat interface has traditionally depended on users copying information into a prompt, describing where files live, and manually carrying results into the next application. Desktop automation can reduce those handoffs, provided users grant the relevant access and organizations establish appropriate controls. 对于开发者和知识工作者来说，这一潜在的转变意义深远。传统的聊天界面依赖于用户将信息复制到提示词中、描述文件位置，并手动将结果带入下一个应用程序。桌面自动化可以减少这些交接工作，前提是用户授予相关访问权限，且组织建立适当的控制措施。
From answers to work across a computer
从回答问题到跨计算机工作
ChatGPT Work is positioned as an agentic layer for work that spans apps and files. OpenAI says the desktop agent can act locally in the background, including interacting with applications, files and browser content. It also highlights plugins, workflows and Scheduled Tasks as ways to connect tools and automate recurring actions across connected apps and local files. ChatGPT Work 被定位为跨应用程序和文件的工作智能体层。OpenAI 表示，桌面智能体可以在后台本地运行，包括与应用程序、文件和浏览器内容进行交互。它还强调了插件、工作流和计划任务，作为连接工具并自动化跨连接应用和本地文件重复操作的方式。
That does not mean every task should be delegated without review. The practical value depends on how clearly a workflow can be defined, the permissions it requires, and the consequences of an incorrect action. For example, moving or modifying local files is fundamentally different from drafting a response in a chat window. The relevant question for teams is not simply whether the agent can perform a sequence of actions, but where human approval should remain part of that sequence. 这并不意味着每项任务都应在未经审查的情况下委派。其实际价值取决于工作流定义的清晰程度、所需的权限以及错误操作的后果。例如，移动或修改本地文件与在聊天窗口中起草回复有着本质区别。团队需要考虑的问题不仅仅是智能体能否执行一系列操作，而是人类审批应在这一序列的哪个环节介入。
| Workflow element | ChatGPT interaction | ChatGPT Work desktop capability |
|---|---|---|
| Primary role | Conversation and responses | Work across apps, files and browser content |
| Local computer interaction | Not the core chat-based interaction model | Can click, type and move files within local applications |
| Contextual work environment | Chat context | Desktop app includes a built-in browser for web-contextual work |
| Automation | User-driven prompts and follow-up | Plugins, workflows and Scheduled Tasks can support connected and recurring work |
| 工作流要素 | ChatGPT 交互 | ChatGPT Work 桌面能力 |
|---|---|---|
| 主要角色 | 对话与回复 | 跨应用程序、文件和浏览器内容工作 |
| 本地计算机交互 | 非核心聊天交互模型 | 可在本地应用程序内点击、输入和移动文件 |
| 情境化工作环境 | 聊天上下文 | 桌面应用包含用于网页情境工作的内置浏览器 |
| 自动化 | 用户驱动的提示与后续操作 | 插件、工作流和计划任务可支持连接与重复性工作 |
For software teams, this may change the shape of routine work rather than eliminate the need for expertise. A developer could use the desktop environment to work through files, browser-based documentation and local applications in one context. But reliable use still requires clear task boundaries, review practices and access decisions. OpenAI says Codex is merging into the new ChatGPT desktop app, bringing coding-oriented work into the broader Chat, Work and Codex desktop experience. 对于软件团队而言，这可能会改变日常工作的形式，而不是消除对专业知识的需求。开发者可以在一个情境下使用桌面环境处理文件、基于浏览器的文档和本地应用程序。但可靠的使用仍然需要明确的任务边界、审查实践和访问决策。OpenAI 表示，Codex 正在合并到新的 ChatGPT 桌面应用中，将面向代码的工作引入更广泛的 Chat、Work 和 Codex 桌面体验中。
Memory makes personalization more consequential
记忆功能使个性化更具影响力
OpenAI’s broader memory and personalization documentation provides the important context for claims that ChatGPT can remember activity across apps and websites. Memory can reference past chats and information accessed through apps to tailor future responses, subject to the user’s memory and data-control settings. The precise controls and data-use details can vary by plan and region. OpenAI 更广泛的记忆和个性化文档为“ChatGPT 可以记住跨应用和网站的活动”这一说法提供了重要的背景。记忆功能可以参考过去的聊天记录和通过应用程序访问的信息，以定制未来的回复，具体取决于用户的记忆和数据控制设置。精确的控制和数据使用细节可能因计划和地区而异。
This distinction matters. Desktop access, agent actions and personalization are related capabilities, but they are not interchangeable. An agent can use a computer to perform a task; memory can make later interactions more tailored by retaining relevant context where settings permit. Organizations evaluating the combined experience should determine what information can be accessed, what can be retained, and which workflows warrant persistent context at all. 这种区别很重要。桌面访问、智能体操作和个性化是相关的功能，但它们不能互换。智能体可以使用计算机执行任务；在设置允许的情况下，记忆功能可以通过保留相关上下文使后续交互更加个性化。评估这种组合体验的组织应确定可以访问哪些信息、可以保留哪些信息，以及哪些工作流确实需要持久的上下文。
The most immediate implications include: 最直接的影响包括：
- Less repetitive prompting when relevant work context is available across interactions. 当跨交互提供相关工作上下文时，减少重复提示。
- More connected workflows across local applications, files and browser-based tasks. 跨本地应用程序、文件和基于浏览器的任务实现更紧密的工作流。
- Higher permission stakes because an agent that can act on a computer may affect business data and operational processes. 更高的权限风险，因为能够在计算机上操作的智能体可能会影响业务数据和运营流程。
- A stronger governance requirement for teams connecting tools or enabling recurring actions. 对于连接工具或启用重复操作的团队，有更强的治理要求。
Governance becomes part of product adoption
治理成为产品采用的一部分
OpenAI frames ChatGPT Work for organizations with enterprise governance, administrative controls and compliance visibility. Its stated controls are intended to help organizations manage who can access what, which tools can be connected and what actions an agent can take. OpenAI also points to the Compliance API as part of this enterprise-oriented governance approach. Those controls are central to adoption, not an administrative afterthought. OpenAI 将 ChatGPT Work 定位为面向具有企业治理、管理控制和合规可见性的组织。其声明的控制措施旨在帮助组织管理谁可以访问什么、可以连接哪些工具以及智能体可以执行哪些操作。OpenAI 还指出 Compliance API 是这种面向企业的治理方法的一部分。这些控制措施是采用的核心，而不是行政上的事后补救。
Cross-application automation can expose a wider set of operational context than a standalone chat session. Teams should therefore map connected applications, identify sensitive file locations, define approval points for consequential actions, and decide which employee roles need access. 跨应用程序自动化可以比独立的聊天会话暴露更广泛的操作上下文。因此，团队应梳理连接的应用程序，识别敏感文件位置，定义关键操作的审批点，并决定哪些员工角色需要访问权限。
The rollout covers Windows and macOS desktops, while ChatGPT Work is also described across web and mobile. OpenAI says the desktop app provides Chat, Work and Codex on all plans, including Free, although access to particular capabilities, controls and personalization behavior may vary by plan and region. Existing Codex users have a path to transition to the updated desktop application. 此次发布涵盖 Windows 和 macOS 桌面，同时 ChatGPT Work 也被描述为支持网页和移动端。OpenAI 表示，桌面应用程序在所有计划（包括免费版）中都提供 Chat、Work 和 Codex 功能，尽管对特定功能、控制和个性化行为的访问权限可能因计划和地区而异。现有的 Codex 用户有途径过渡到更新后的桌面应用程序。
For businesses, desktop agents are most useful when they are introduced as governed workflow components rather than universal autonomous operators. Scalevise can help teams assess high-value use cases, define permissions and human-review points, and connect AI capabilities to real operating processes through AI consultancy for governed desktop-agent adoption. 对于企业而言，桌面智能体作为受治理的工作流组件引入时最为有用，而不是作为通用的自主操作员。Scalevise 可以通过 AI 咨询帮助团队评估高价值用例、定义权限和人工审查点，并将 AI 能力连接到实际运营流程中，以实现受治理的桌面智能体采用。