Abstract: Despite the milestone progress of artificial intelligence (AI) in recent years, significant bottlenecks persist in achieving general intelligence, particularly in energy-efficient computing, causal reasoning, cross-task generalization, and continual learning. These persistent limitations mandate shifting research towards the biological brain, seeking Neuro-Inspiration and Intelligence co-evolution to develop computing systems with superior energy efficiency and generalization. The “AI-Brain Co-evolution” forum established a core consensus: Significant disciplinary barriers persist among the understanding of neuroscientific mechanisms, the design of intelligent algorithms, and the implementation of computing systems. Consequently, neuro-inspired mechanisms have proven difficult to translate into executable intelligent algorithms and practical systems. The forum defined key pathways for synergy. Intention: Research must return to the essential understanding of neural mechanisms, merging neuro-inspiration with system implementation. Pathway: Future AI must overcome structural homogeneity and reliance on global backpropagation by adopting the brain’s spatio-temporal heterogeneous architecture and localized learning mechanisms. Validation: Brain-Computer Interfaces (BCIs) and Neuromedicine provide advanced validation modalities via physical intervention, accelerating the transition from correlation to causality. Assurance: A problem-oriented, interdisciplinary mechanism is required to support long-term exploration and differentiated contributions.