Scientists from Loughborough University have developed a mathematical blueprint for transparent artificial intelligence, potentially ending the era of opaque black box systems. Published in Physica D: Nonlinear Phenomena, the study outlines a prototype architecture that reveals how machines learn, remember, and make decisions. Unlike conventional AI, which often obscures its internal logic, this new system is designed with built-in transparency, allowing its cognitive processes to be fully traced and understood. The prototype features a distinct structure comprising both a processing unit and a memory system. It can learn continuously without suffering from catastrophic forgetting, a common issue where AI models lose previously acquired knowledge when learning new tasks. The system also mimics human cognitive traits, such as strengthening or forgetting information over time, while avoiding the formation of false memories. In early tests, the model successfully learned musical notes and phrases without supervision and identified colors from cartoon images. Throughout these tasks, its behavior remained predictable and traceable. Lead author Dr. Natalia Janson of the Department of Mathematical Sciences emphasized the shift in philosophy behind the project. She stated that intelligence has historically been treated as an emergent property hidden inside a black box. The team's goal was to rethink AI from the ground up to create a system where the inner workings of cognition are completely visible. A key element of this approach is the use of a mathematical concept known as a plastic vector field. This framework models how information changes over time in a way that mirrors how the human brain processes and stores data. By applying this concept, the researchers ensure that every stage of learning and cognition is tracked from the outset, rather than trying to add explainability to an existing system later. Professor Alexander Balanov of the Department of Physics noted that the limitations of current artificial neural networks are often rooted in their fundamental design. He explained that their architecture makes it impossible to fully control how they learn and store information, which inherently hinders explainability. The new approach aims to solve these root causes by addressing the connection between memory, behavior, and physical structure. While the prototype demonstrates significant promise, it remains a relatively simple system that requires scaling for real-world application. The Loughborough team plans to further develop the technology and explore its integration into new hardware types. Their ultimate objective is to create AI that is not only powerful but also trustworthy and easy to understand. Professor Balanov expressed confidence that this research will bring society closer to reliable technologies, ranging from safer healthcare tools to more accountable automated decision-making systems.