OpenAI Launches Dots, Always-On AI Agents Inside ChatGPT

The company describes Dots as more autonomous than existing agentic tools. Users set goals and boundaries, and the Dots determine how to work toward them. Each Dot has its own cloud computer and browser, allowing it to interact with the web, use apps, run code, and work with other plugins. That setup keeps the user's computer separate and limits the potential damage an autonomous agent could cause, TechRadar reported.

A Dot's activity appears as in progress, scheduled, or completed. Possible uses include checking a user's calendar to produce a morning report or reviewing information from connected services and apps. OpenAI has also added Dots to Slack and Teams, with text support planned soon, according to the report.

Built-in safeguards are intended to pause or stop a Dot's work if the system detects malicious instructions or behavior. TechRadar reported that Dots are initially available to Pro and Business Premium users at no additional cost. Only work performed toward a set goal consumes usage limits, while conversations with Dots are free.

Users must set up Dots on the desktop app or a browser first. After that, they can interact with Dots on the mobile app. OpenAI did not share plans for a broader rollout. TechRadar noted that, based on OpenAI's previous deployment phases, Dots could extend to other subscription types in the future once early users understand real-world use cases and costs.