Recent advances in artificial intelligence, particularly through reinforcement learning, deep neural networks and large language models, have revolutionised automated pronunciation evaluation systems.