Abstract
The hermaphroditic Caenorhabditis elegans, with its fully mapped connectome of 302 neurons, offers a paradigmatic example of how a minimal nervous system governs biotic, adaptive, and context-dependent behaviors. In contrast, modern artificial intelligence systems achieve intelligence through scale rather than efficiency, relying instead on massive datasets and artificially engineered architectures. This mini-review explores how Caenorhabditis elegans neural circuits can inform the development of more efficient and flexible artificial neural networks. We highlight recent studies that translate the principles inherent to Caenorhabditis elegans neural circuits into artificial neural network architectures, with applications in machine control and image classification, resulting in enhanced robustness and improved performance. By distilling neural principles from the simplest known nervous system, this mini-review outlines a pathway toward compact, adaptive, and biologically inspired artificial intelligence systems.