Embodied AI and industrial automation trends highlighted

At the ISPE event ‘Robotics & AI and RAYA’ held at Takeda, Professor Péter Galambos of Obuda University discussed the current state and challenges of embodied artificial intelligence. While AI has seen massive shifts through deep learning and large language models, embodied AI continues to face the Moravec paradox, where digital intelligence excels but physical manipulation remains difficult due to the complexity of high-dimensional sensor streams.

Research in controlling physical agents currently relies heavily on reinforcement learning and imitation learning. In parallel, the automation industry is looking toward software innovation through the ‘automation app award 2027’. This award seeks market-ready applications in categories such as engineering, robotics, production, and AI-driven software-defined automation. Applications for the award open on October 12, 2026, with a submission deadline of December 31, 2026.