OpenAI's Dots Agent Wants to Run Your Life

OpenAI has introduced Dots, an always-on AI agent designed to automate online tasks such as buying furniture and managing daily errands. A WIRED reporter tested the agent and found it buggy, including a failure to solve a CAPTCHA, but the experience still highlighted OpenAI's ambition to build persistent personal agents. The launch matters because it moves AI beyond chat-based assistance toward autonomous action across the web. Instead of simply answering questions, an agent like Dots is meant to carry out multi-step tasks, interact with websites, and make decisions on a user's behalf. That shift could save time and reduce friction for routine chores. It also raises major questions about reliability, privacy, security, and user control. If an agent can log into accounts, enter payment details, or submit forms, mistakes can have real consequences. CAPTCHAs and anti-bot systems are only one obstacle; agents must also handle ambiguous instructions, changing websites, and sensitive data. The WIRED test suggests Dots is not yet polished, which is unsurprising for an early product. But the direction is significant. OpenAI and other companies are racing to create AI that acts continuously in the background, not just when prompted. That could eventually make personal agents feel like an operating layer for daily life. For users, the appeal will depend on trust: the agent must be transparent, controllable, and secure enough to delegate meaningful tasks. For regulators and platforms, always-on agents may force new rules about automation, consent, and accountability. Dots is therefore less a finished assistant than a preview of a coming wave, one where AI does not just talk, but tries to get things done.