Traditional experimental teaching content for robotics courses is primarily confined to the discipline of robotics itself， lacking sufficient innovation and failing to meet the high standards for innovative capabilities demanded of emerging engineering professionals in the context of developing new types of productivity. In response to these issues， artificial intelligence technologies such as deep learning and reinforcement learning have been deeply integrated into the field of robotics to design and develop comprehensive and interdisciplinary teaching experiments on "AI + Robotics". The experiments enable users to manipulate robots to autonomously grasp target objects through gestures， integrating five "AI+" functional modules： gesture recognition， localization and mapping， path planning， target object detection， and robotic arm grasping. This robotic teaching experiments demonstrate a high level of intelligence and autonomous decision-making capabilities， fostering students'' abilities in interdisciplinary application and innovation. It provides teaching practice support and innovative impetus for the development of new productivity in the robotics industry.