Abstract: In 2024, Hopfield and Hinton were awarded the Nobel Prize in Physics for their fundamental contributions to neural-network based artificial intelligence (AI), which triggered much debate on whether AI belongs to physics and how physics contributes to AI. Here, we review the history of how physics contributed to the early development of AI, and in particular we highlight two important scientific branches that were initiated by applying physical concepts to study neural networks. We also discuss future directions for understanding and improving AI, and for resolving the nature of intelligence.