AI TRENDS | World Labs and Li Feifei Critique Misuse of 'World Model' in AI

In June 2026, World Labs, in collaboration with Li Feifei, released a conceptual article critiquing the term 'world model' as one of the most overused in the AI field. According to PANews, the article highlights the lack of commonality among various AI systems such as video generators, autonomous driving prediction networks, and robotic control models, despite all being labeled as 'world models.' Instead of introducing a new model, the article revisits the origins of the POMDP theory, categorizing all systems into three functional projections: renderers, simulators, and planners. This approach aims to cut through marketing jargon and provide a clearer understanding of the true capabilities and limitations of different AI systems.