Independent tech intelligence, checked against primary sources.

TechPulseMind Useful technology.
No manufactured hype.

OpenAI Dots Bring Always-On AI Agents to ChatGPT Users

OpenAI's new dots are always-on agents with their own cloud computer, cross-app connections and controls for approvals and sensitive actions.

OpenAI Dots Bring Always-On AI Agents to ChatGPT Users

OpenAI has introduced dots, a new kind of always-on AI agent designed to keep working across projects instead of waiting for a fresh prompt every time. Powered by GPT-6 Astra, each dot gets its own cloud computer and browser, can connect to supported apps, and can continue work in the background under rules set by the user.

The launch moves ChatGPT further from one-off conversations toward delegated work. A dot can follow a project over time, learn from feedback and bring completed work or decisions back for review. OpenAI is beginning the rollout for Pro and Business Premium users in eligible markets, while Enterprise users, including Edu and Healthcare organizations, can try the beta when a workspace administrator enables it.

What an OpenAI dot can do

A dot can work on several projects without forcing the user to manage a separate conversation for every task. OpenAI says dots can use their own browser and cloud computer, and connect to more than 4,000 apps through the company’s plugin ecosystem. Users can inspect the dot’s computer while it works and can choose whether to connect another device, including a laptop.

OpenAI’s launch examples show a dot investigating recurring product feedback and preparing tested fixes, revising launch materials after requirements change, updating an analysis when new evidence arrives, updating a proposal as requirements shift, and preparing content assets for approval. These are examples supplied by OpenAI, not independent performance tests.

Communication is not limited to one ChatGPT window. OpenAI says users can message a dot in ChatGPT on desktop, web and mobile. Slack and Microsoft Teams are supported communication channels, and text messaging is planned.

Background work comes with controls

The most important difference between dots and a conventional chatbot is that a dot can continue working and prepare actions. OpenAI pairs that autonomy with several controls. Users choose which apps a dot can access, can inspect progress in Activity View and can redirect the work while it is running.

For background “proactive research,” OpenAI says dots use already-connected apps through tools restricted to read-only access. Those tools cannot send messages, change app content or control the user’s browser or computer. Broader actions are governed by built-in rules and optional Custom Rules, which can allow a specific action, require approval or block it.

An auto-review system checks proposed actions against the user’s instructions, Custom Rules and safety requirements. Some work can proceed, some must wait for approval and some remains reserved for the user. OpenAI gives changing a password as an example of a sensitive task that a dot cannot complete on the user’s behalf.

Privacy and availability

Each dot normally operates on its own cloud computer, separate from the user’s device unless the user explicitly connects one. OpenAI also says supported website credentials can be used without exposing saved passwords to the model.

Content from ChatGPT Business, Enterprise and Edu workspaces is not used to improve OpenAI’s models by default. On personal plans, the relevant training choice remains under the user’s data controls. OpenAI says proactive research and a dot’s internal notes are not used directly for training, although information from them may influence an eligible conversation or task depending on the account settings.

The first dot is included with Pro or Business Premium at no additional charge, according to the launch announcement. To begin, eligible users create a dot from the ChatGPT desktop app or a desktop browser, connect the apps they want it to use and complete the introduction flow. After setup, the same dot can also be reached from the ChatGPT mobile app.

OpenAI explicitly warns that dots can still make mistakes. The useful question is therefore not whether a dot can run continuously, but which permissions it receives, which actions require approval and how carefully consequential output is reviewed.

Source

Official announcement or documentation

Some links on this page may be affiliate links. If you buy through them we may earn a commission at no extra cost to you. See our affiliate disclosure.