OpenAI‘s collaboration with Figure has brought to life a humanoid robot, Figure 01, that not only interacts with humans but also carries out tasks with an unprecedented level of autonomy and intelligence.

Developed by Figure, a company born from the minds of experts formerly associated with giants like Boston Dynamics and Google DeepMind, this humanoid robot is powered by OpenAI’s sophisticated AI systems. It can perform tasks that were once considered exclusively human, such as handing out food, picking up trash, and organizing dishes.

The robot’s demonstration, as shared by Brett Adcock, Figure’s co-founder and CEO, on social media, showcases its ability to interact with humans and understand commands. In one instance, Figure 01 precisely identifies and hands over an apple, following a human’s request, demonstrating its impressive visual and auditory processing skills.

OpenAI’s large vision-language model (VLM) allows the robot to process visual data and auditory commands simultaneously, enabling it to understand and respond to complex instructions. Brett Adcock elaborated on the robot’s functionality, noting that the entire operation is conducted through end-to-end neural networks, with no teleoperation, highlighting the autonomous nature of the robot’s actions.

Corey Lynch, the AI director at Figure, emphasized that the robot’s behaviors are learned and executed in real-time, allowing it to interact with the environment and people in a natural and intuitive manner. The integration of OpenAI’s AI model equips Figure 01 with the ability to not only comprehend instructions but also to reason and make decisions independently.

Adcock reassures that Figure’s vision is to enhance human capabilities and improve life quality, explicitly stating the company’s commitment to avoiding the deployment of their robots in military or harmful applications. The long-term goal is to develop robots that can safely and effectively coexist with humans, contributing positively to society and the economy.