Doosan Robotics has partnered with Microsoft and Doosan Digital Innovation to develop a context-aware control system for collaborative robots using generative AI. Under a new memorandum of understanding, Doosan will leverage Microsoft's Azure OpenAI to integrate the GPT language model into its cobots.

The collaboration sees Azure providing access to GPT, while Doosan Digital Innovation adapts the model for robotics applications. Doosan aims to shorten programming times and expand capabilities through GPT's comprehension skills. As a large language model trained on vast data, GPT can understand and generate human-like text.

By integrating GPT, Doosan seeks to enable cobots to self-correct during operation, anticipate tasks, and adjust actions based on context. For example, users could voice command meal preparation, with the robot optimizing the steps autonomously. GPT could also allow robots to learn from existing code, eliminating programming repetitive new tasks.

Doosan plans initial food and beverage industry testing this year before potentially expanding to manufacturing. This partnership exemplifies growing interest in using generative AI to transform next-generation robotics.

In fact, at October's RoboBusiness event in Santa Clara, CA, robotics leaders will keynote on applying large language models like GPT to design, training, simulation, algorithms, and commercialization. Panelists include Agility Robotics' Pras Velagapudi, Formant's Jeff Linnell, UC Berkeley's Ken Goldberg, NVIDIA's Amit Goel, and OLogic's Ted Larson.

By leveraging AI advancements from Microsoft, Doosan Robotics aims to pioneer more intuitive, context-aware cobots. Early adoption for food preparation could later disrupt wider automation markets as capabilities progress. With GPT integration, Doosan is positioning itself at the forefront of AI robotics innovation.