China Releases Embodied AI Version of Smart Crop Breeding Robot 'Ji'er'

On September 4 in Beijing, researchers unveiled an upgraded embodied AI version of 'Ji'er,' an autonomous crop breeding robot capable of flower stamen recognition, navigation, precision cross-pollination, and phenotypic data collection. Developed by a team led by Xu Cao at the Institute of Genetics and Developmental Biology under the Chinese Academy of Sciences, the robot works alongside bio-engineered crops designed to expose flower stigmas for robotic access. This development represents a significant cross-disciplinary convergence of biotechnology and embodied AI in precision agriculture. By automating labor-intensive manual cross-pollination and field phenotyping, the technology could drastically accelerate crop breeding cycles and lower labor costs in hybrid seed production. During demonstrations, 'Ji'er' used dual-arm collaboration—holding a pollen tube in one hand and a pollination pen in the other—to complete a single tomato stamen pollination in under 10 seconds. The system navigates dense foliage to pinpoint needle-tip-sized stamens, with a single pollen dip supporting approximately 50 individual pollination tasks.