OpenAI tries disarming AI angst with cute graphics and always-on agents

Your AI assistant can now work for you around the clock without eating into your chat credits.

OpenAI has launched "dots," a new feature that offers always-on AI agents designed to be user-friendly and helpful. These "dots" are persistent compute tasks with a cute, anthropomorphic visual representation, aimed at disarming anxiety around AI. CEO Sam Altman described them as AI helpers that "always have your back," capable of performing long-running tasks 24/7 and retaining context. Dots use OpenAI's GPT-6 Astra model and can interact with approximately 4,000 apps via connectors. While initially appearing to be included with existing subscriptions without additional cost, OpenAI clarified that dots are part of Pro or Business Premium plans and may incur future flat monthly fees for enhanced capabilities or more dots. Conversations with dots will not count against ChatGPT usage limits, but tasks initiated by dots within applications like Codex or ChatGPT Work will still be subject to standard usage allowances. OpenAI also introduced ChatGPT Space, a collaborative workspace for humans and agents, and the Decisions API for faster AI responses.

The introduction of "dots" signals a strategic move by OpenAI to make AI agents more accessible and less intimidating, directly addressing public apprehension. By embedding these agents into user interfaces with playful graphics and framing them as helpful companions, OpenAI aims to foster a more positive user perception and drive adoption. This approach positions OpenAI not just as a provider of AI models but as a platform for integrated, persistent AI assistance, blurring the lines between traditional software and AI-driven services. The company's emphasis on user experience and a "Renaissance" rather than an "industrial revolution" framing suggests a long-term strategy to democratize advanced AI capabilities.

OpenAI's strategy with "dots" and ChatGPT Spaces points toward a future where AI agents are deeply integrated into daily workflows, acting as persistent collaborators rather than standalone tools. The pricing model, which initially offers dots without extra conversational cost but hints at future paid enhancements, suggests a tiered approach to AI agent utility. This could lead to a market where users pay for the depth and breadth of agentic capabilities they require, influencing how individuals and businesses budget for AI productivity tools. The accompanying developer tools and APIs also indicate OpenAI's intent to build out a robust ecosystem around these persistent AI agents, encouraging third-party integration and innovation.