OpenAI announced its new AI agents, named Dots, at its annual DevDay conference on Tuesday. These agents are designed to function as continuous AI assistants, capable of working across various connected applications in the background and learning user preferences over time. The company stated that Dots are powered by its GPT-6 Astra model and utilize their own cloud computer to access a web browser and more than 4,000 supported applications. Users can interact with a Dot through interfaces such as ChatGPT, Slack, and Microsoft Teams, with text messaging capabilities planned for future release.
The introduction of Dots follows Meta's launch of its own AI agent, Muse, earlier this month. Meta's Muse, powered by its Muse Spark model, also operates as a personal AI agent that can proactively assist with goals and suggest ideas. Muse runs on a dedicated secure virtual machine and can work across daily applications, learning from interactions. Both Dots and Muse represent a push by major technology companies into the market for persistent, agentic AI.
OpenAI CEO Sam Altman described Dots as an "AI helper that always has your back," likening them to the AI companions often depicted in science fiction. He emphasized that Dots aim to take on both personal and professional tasks proactively, allowing users to delegate complex work. Internal examples of Dots' use at OpenAI include automatically investigating bugs reported in Slack and generating invoices for user approval. Dots can also conduct "proactive research" using read-only connections to connected tools when not actively engaged in a specific task.
Dots are initially available to ChatGPT Pro subscribers and Business Premium users in eligible markets. OpenAI plans to expand availability in the future. The company has stated that it does not use business customers' data for AI training by default. Each Dot operates with built-in rules for independent action, but users can establish custom boundaries for when agents require permission, supported by auto-review features to ensure compliance with safety requirements. OpenAI also introduced ChatGPT Space, a new collaborative platform where human teams, ChatGPT, Codex, and Dots can work together within shared documents.