Abstract: Promoting artificial intelligence (AI) technology application to facilitate revolutionary leaps in productivity holds strategic importance for accelerating the formation of new-type enterprise labor structures characterized by human-machine collaboration,cross-boundary integration,and co-creation sharing.However,at the micro-firm level,critical questions remain unresolved:Can AI technology application promote enterprise digital-intelligent innovation? Through which mechanisms does it enhance such innovation,and is labor skill structure optimization the key mechanism? What are the internal mechanisms driving enterprise labor skill structure optimization? Based on this research gap,this study selects Chinese A-share listed manufacturing firms (2015—2024) as research samples and employs corporate annual reports to construct firm-level AI technology application indicators through: ①Preprocessing annual reports crawled from Sina Finance,incorporating the AI dictionary into Jieba library for word segmentation. ②Adopting the Skip-gram model to screen semantically closest words based on cosine similarity. ③Referencing Stanford HAI's 2025 AI Index Report and employing Word2vec semantic expansion to generate an AI dictionary comprising 100 terms. ④Measuring firm-level AI technology application by calculating the natural logarithm of one plus AI keyword counts.Concurrently,based on IncoPat patent database,this study measures enterprise digital-intelligent innovation output by classifying invention patents according to technical efficacy trends. Our empirical findings reveal:①AI technology application exerts a significantly positive impact on enterprise digital-intelligent innovation,remaining robust after endogeneity treatments and robustness checks. ②AI technology application enhances digital-intelligent innovation by optimizing internal labor skill structures,improving technology affordances,and increasing knowledge diversity. ③Technology affordances and knowledge diversity constitute key mechanisms driving labor skill structure optimization,significantly incentivizing firms to increase demand for non-routine high-skilled labor.Specifically,interactive affordances demonstrates stronger positive effects than autonomous affordances; unrelated knowledge diversity exhibits stronger positive impacts than related knowledge diversity. ④Heterogeneity analysis indicates AI technology application shows more pronounced effects in non-state-owned enterprises,proves effective exclusively for large-scale enterprises,and capital-intensive industries exhibit higher sensitivity compared with labor-intensive and technology-intensive industries. This study's marginal contributions include:①Employing MD&A textual data to capture autonomous and interactive affordances characteristics,and disaggregating knowledge diversity into unrelated and related dimensions based on patent information knowledge structures,thereby examining how different technology affordances and knowledge diversity promote labor skill structure optimization. ②Using neural network models and machine learning methods to construct firm-level AI technology application indicators based on annual reports. ③Classifying patent applications based on technical efficacy to measure “digitalization” and “intelligentization” capabilities,overcoming limitations of traditional methods focusing solely on innovation scale while neglecting quality characteristics. Based on these findings,this study advances policy recommendations:First,guiding enterprises toward integrated collaborative AI development while strengthening intelligent upgrades of research platforms; Second,encouraging enterprises to leverage AI's role in creating new positions while reinforcing non-routine high-skilled talent creativity; Third,fully exploiting autonomous and interactive affordances to promote manufacturing transformation; Fourth,directing enterprises to apply AI technology to transcend traditional search limitations,enhancing cross-module knowledge matching efficiency.This research contributes novel empirical evidence for understanding mechanisms through which AI technology application empowers enterprise digital-intelligent innovation from the micro-firm perspective.