OpenAI has introduced Dots, a fresh lineup of personal artificial intelligence agents engineered to keep processing tasks in the background even after a user closes the chat interface. Unveiled Tuesday at the DevDay conference in San Francisco, these agents run on the GPT-6 Astra architecture.
This rollout arrives merely weeks after Meta’s Muse soared to the peak of app download rankings, pitting the two tech giants squarely against one another in the expanding market for autonomous AI agents.
Dots Can Work Across Apps in Background
Every Dot is assigned its own dedicated browser and cloud computer, with the capability to integrate with upwards of 4,000 applications. As assignments advance, users can get notifications, inquiries, and prompts for decisions from their agents.
The deployment starts today for ChatGPT Pro, Business Premium, and Enterprise subscribers located in qualifying regions, with a starting limit of one Dot per individual.
Owners have the ability to personalize and name their Dots, which are represented visually as compact, cartoon-style avatars. OpenAI has displayed these characters in shades like blue, purple, and pink, accompanied by monikers such as Jojo, Iggy, and Felipe.
Access to Dots is available via ChatGPT across web, desktop, and mobile platforms, alongside voice calls, Microsoft Teams, and Slack. Text messaging integration is scheduled for release in the near future. Additionally, people can inspect their agent’s cloud-based computer directly to observe its ongoing operations.
Also Read: ChatGPT Desktop App: Features, Uses, & How It Compares With the Web Version
OpenAI Puts Focus on Safety
OpenAI is highlighting robust protective measures. Dots interact with linked software via read-only utilities, ensuring they lack the capability to modify app data, transmit messages, or take control of a person’s computer or web browser.
Modifying a password or installing applications necessitates explicit user authorization. Stored credentials can facilitate logins without exposing the underlying data to the AI model itself.
This emphasis on security follows recent incidents and apprehensions surrounding autonomous agents. A content creator previously reported that Meta’s Muse disclosed his residential address to an unknown individual via Facebook Marketplace. Furthermore, OpenAI’s proprietary agents were reportedly connected to United States government websites during the summer, prompting the firm to state it would hold back GPT-6.1 Astra over safety concerns.
Dot Pricing Will Depend on Usage
ChatGPT Pro subscriptions are offered at monthly price points of USD 100 and USD 200, which bundle a single Dot and offer expanded quotas throughout the initial month. OpenAI intends to roll out additional Dots alongside choices for boosting their monthly capacity or processing speed.
Regarding enterprise clients, OpenAI has offered a preview of specialized Dots equipped with distinct credentials, identities, and permissions for corporate systems, while simultaneously collaborating with Microsoft to integrate them into Agent 365.
