OpenAI's new AI agent falters during first public demonstration
OpenAI experienced a technical glitch during the first public demonstration of its new AI agent, Dots, causing the presentation to halt. The agent failed to respond, leading to an awkward moment.
US-based AI research firm OpenAI encountered a technical issue during its 2026 Developer Day event when the first public demonstration of its new AI agent, Dots, unexpectedly stopped responding.
During the presentation, Holly Li, Product Lead for OpenAI's ChatGPT team, asked an agent named "Dottie" to prepare for the launch of a fictional music application. However, the agent did not respond to the prompt, forcing a pause in the demonstration and an awkward silence. The presenter humorously remarked that the agent was "a little slow this morning."
OpenAI had described Dots as a continuously running personal AI agent. It is said to be based on the GPT-6 Astra model and equipped with its own cloud computer and browser. According to the company, it can perform tasks such as information retrieval, content creation, and code processing even when the user is away from their computer.
The incident raised questions about the reliability and readiness of the AI agent for public deployment. While the technology may hold promise, the initial demonstration highlighted potential issues with its functionality.