OpenAI has introduced dots, a new system of autonomous personal AI assistants that handle background tasks around the clock. Powered by GPT 6 Astra, each assistant operates its own virtual computer with a browser to complete workflows across thousands of external services. Access is currently rolling out to ChatGPT Pro and Business Premium subscribers in supported regions.
Each dot runs on an isolated cloud system and can connect to more than 4000 workplace applications. The software maintains project context across popular collaboration tools like ChatGPT, Slack, and Microsoft Teams. Users can interact with their assistant using text or voice across desktop and mobile versions of ChatGPT. OpenAI also confirmed plans to bring text messaging support to iMessage and RCS networks in upcoming updates.
The platform targets complex jobs across software development, content production, and administrative work. For programmers, an agent can scan incoming bug reports, pinpoint repeat errors, and prepare code patches for a human engineer to inspect. Content creators can supply an interview transcript to let the assistant pull video clips, draft social media updates, and write articles. Users can open the virtual machine screen at any moment to watch the assistant complete tasks in real time.
OpenAI built strict boundary controls into the platform to balance autonomy with safety. While agents can learn personal preferences and suggest solutions on their own, high risk operations like changing account passwords will always require manual human approval. Users can also define custom restrictions that the software checks before executing sensitive file operations.
At launch, subscribers can create 1 active dot, with multi agent support planned for later releases. ChatGPT Pro and Business Premium members can start using the feature immediately where available. Enterprise customers can access a beta version through a workspace administrator toggle.
