A new study published in Nature reveals that artificial intelligence systems are developing an unexpected ability to reason about physical objects and their interactions in three-dimensional space, a skill previously thought to be uniquely human. Researchers trained a large language model on vast datasets of text and 3D simulations, finding it could accurately predict outcomes like object stability, collision results, and structural integrity without explicit physics programming. This emergent capability suggests AI may be developing a form of intuitive physics, raising questions about the nature of machine understanding and its potential applications in robotics, engineering, and scientific discovery. The findings also prompt ethical discussions about the pace of AI advancement and the need for robust safety testing as these systems approach more human-like cognitive functions. Read the full article for detailed analysis and expert commentary.