OpenAI launches Dots, always-on agents powered by GPT-6 Astra with their own cloud computer, in ChatGPT for Pro, Business Premium, and Enterprise users

Your personal AI agent can now do work for you, like sending invoices or writing code, after you grant it permission.

OpenAI has launched Dots, new proactive AI agents designed to perform tasks autonomously for users. These agents operate with their own cloud computers, browsers, and can integrate with over 4,000 apps via OpenAI's plugin ecosystem. Dots can also interact with external services like Slack and Microsoft Teams. The agents can assist with tasks such as preparing invoices, integrating data into research papers, or collecting user feedback for software development. At launch, Pro, Business Premium, and Enterprise users in select markets will receive one customizable Dot. OpenAI plans to enable users to control multiple Dots in the future. Safety features include password management without model exposure and read-only tool restrictions for proactive research. Dots are rolling out to eligible subscribers, with the first Dot offered for free.

Dots represent a significant step towards more autonomous AI agents, moving beyond reactive chatbots to proactive assistants that can manage tasks and interact with a wide range of applications. This development signals a shift in how users will engage with AI, potentially automating many routine digital activities and integrating AI deeper into daily workflows across personal and professional settings. The ability to customize agent behavior and permissions addresses early concerns about AI autonomy and control, while the planned expansion to multiple agents hints at future complex, collaborative AI systems.

The integration of Dots across multiple OpenAI platforms and third-party applications like Slack and Teams suggests a strategy to embed these agents deeply into user ecosystems. For businesses, this could mean streamlined operations and enhanced productivity, but it also raises questions about data privacy, security, and the oversight required for autonomous AI actions. As OpenAI emphasizes safety features, the adoption rate will likely depend on user trust in these safeguards and the demonstrable value proposition of agents handling complex tasks with minimal human intervention.