Social RoboticsFigure 01 holds a spoken conversation while it works
A Figure demo showed its humanoid describing what it saw and acting on requests using an OpenAI vision-language model.
Figure released a video in which its Figure 01 humanoid talked with a person in real time, explained what it could see and handed over an apple when asked. OpenAI's model handled the visual scene and speech understanding, while Figure's own neural networks turned that into fast hand and arm movements. Chief executive Brett Adcock said the robot was not teleoperated and that the footage ran at real speed. The clip became a widely shared example of conversational AI giving a humanoid robot a natural spoken interface.
Source: New Atlas
This archive entry summarises reporting published elsewhere; follow the source link for the full story. The summary was drafted with AI from that reporting.