Chinese robotics companies are currently hampered by insufficient data and an inadequate “brain” to enhance how their machines interact with the real world, industry insiders reported at the World Artificial Intelligence Conference (WAIC) in Shanghai, which wrapped up on Monday.
The primary challenge for the embodied AI sector is to “integrate hardware, data, models and real‑world scenarios into a closed‑loop iterative system”, Wang Xiaogang, co‑founder of SenseTime and chairman of its robotics spin‑off Ace Robotics, told the South China Morning Post on Saturday.
A large volume of training data is currently gathered from human demonstrations. However, for this data to translate into more capable embodied AI, robot hardware design must be optimized together with data‑collection techniques and physical architecture, Wang explained.
He added that deployment of robots in the real‑world environments where data originates remains insufficient.
“The key question is how to unlock these scenarios and replicate them at scale,” he said.
The quantity of multi‑modal data about the physical world is still far short of what is available for large language models, according to Yao Maoqing, partner and senior vice‑president of Shanghai‑based humanoid robot manufacturer AgiBot. This shortage creates a major bottleneck in training the so‑called world models that are intended to enable next‑generation humanoid robots to model and navigate their environment, he said.