The company OpenAI is making strides in the field of robotics. In February 2025, OpenAI launched a robotics laboratory in San Francisco. Since then, the laboratory has expanded more than fourfold and now operates around the clock. About 100 data collectors and at least a dozen robotics engineers remotely control robotic manipulators to generate training data. The primary focus is on cost-effective robotic manipulators rather than fully-fledged humanoid robots. Workers use a 3D-printed controller called GELLO, which directly maps human hand movements to the movements of robotic manipulators. These robots are being trained to perform everyday tasks, such as loading bread into a toaster and folding clothes.
OpenAI’s strategy is based on collecting large volumes of human-created data, similar to how early language models were trained. Experts say the biggest challenge in robotics is not algorithms but high-quality data. A second robotics laboratory is planned in California, indicating OpenAI’s long-term interest in this field. Although creating fully-fledged humanoid robots remains a future goal, OpenAI is steadily laying the groundwork for its achievement.
Recently, OpenAI has also been exploring advancements in AI safety and ethics, ensuring that its robotic technologies align with human values and societal norms. This is particularly relevant given the increasing complexity and autonomy of AI systems. The company’s ongoing research involves collaboration with academic institutions to refine the ethical frameworks that guide AI development, thus fostering a more responsible and beneficial integration of AI into daily life.
