IGNITE SEMINA | World Models, Large Language Models, and the Future of AI

On the afternoon of November 19, 2025, the IGNITE Institute hosted a Seminar presented by Assoc. Prof. Dr. Nguyen Ai Viet, focusing on the new research of Prof. Fei-Fei Li (Stanford & World Labs) regarding world models – an approach regarded as a significant stride in the post-LLM era.

The session introduced and analyzed prominent perspectives within the global AI research community, including:

- Text-based AI is powerful but does not yet “understand” the world.

- The role of perception and sensory awareness (Fei-Fei Li).

- The necessity of world models and autonomy to advance toward Autonomous Intelligence (Yann LeCun).

- The concept of embodiment, positing that intelligence must be tethered to the environment and real-world interaction.

- The assertion that Large Language Models (LLMs) cannot evolve into Artificial General Intelligence (AGI) if standing alone.

From these premises, the Seminar expanded the discussion on future AI trends – where artificial intelligence must transcend text to move closer to human-like capabilities in reasoning, prediction, and action.

The session also raised critical strategic questions for Vietnam: If world models become a global trend, what preparations are required to seize opportunities and position national AI capabilities in the next phase?