Figure Unveils Conversational Humanoid Robot Powered by OpenAI
Robotics startup Figure demonstrates its humanoid robot engaging in real-time conversation and completing tasks using OpenAI's multimodal AI technology.
Robotics startup Figure shares a new video demonstrating its humanoid robot, Figure 01, engaging in a real-time conversation with a human engineer. This breakthrough comes from a recent partnership with OpenAI, which integrates advanced visual and language intelligence directly into the robot's system.
In the demonstration, Figure 01 successfully identifies objects like an apple and dishes, hands over food when asked, and collects trash while answering questions simultaneously. The robot processes camera images and microphone audio through a large multimodal model, allowing it to describe its visual experience, plan future actions, and explain its reasoning out loud.
Senior AI Engineer Corey Lynch emphasizes that all actions in the video run at normal speed and are entirely learned rather than remotely controlled. By applying common sense reasoning to its environment, Figure 01 represents a major step forward in creating robots that understand and react naturally to human interactions.