What happened

At a major global industrial internet conference, Lu Tingjie argued that AI should be more about acting than talking, stressing world models and embodied intelligence. He warned that many embodied AI systems still rely on preprogrammed control and lack true understanding of the physical world.

Why it matters

The claim points to a path where AI could more effectively integrate with traditional industries, but also highlights that current embodied AI has gaps,so gains are not guaranteed for all contexts or people.

In his talk, Lu Tingjie described the shift from mere computation to understanding and acting in the physical world, emphasizing world models that simulate real objects and events. He noted that many embodied systems still depend on fixed programming and human control, which limits autonomy.

The potential is a more responsive AI that can adapt to industrial settings,from engines to traffic scenarios,by modeling digital twins and autonomous decision-making. Yet he also warned that these advances must be carefully developed to avoid overestimating current capabilities and to account for infrastructure and safety limits.

What this does not tell us

Source is a single conference presentation; observations may reflect promotional framing and do not establish universal outcomes.

FOR PEOPLE

No direction stated

A practical view helps people learn AI without hype or fear.

FOR AI AND ITS OPERATORS

Benefits reported

AI enables industrial agents and world-model-based control infrastructure

These are two separate readings of what the sources describe. Reported claims and risks do not by themselves establish a real-world effect.

Original sources · 1
  1. Lü Tingjie: Let artificial intelligence “act rather than talk” – full-space intelligence and global models empower the development of industrial AI agents ↗Xinhua Net · 2026-09-09

Reporting discovered in China. Discovery market does not mean the event happened there.