Artificial Intelligence Technology, Organizational Learning Capability, and Corporate Innovation Performance: Evidence from Chinese Specialized, Refined, Unique, and Innovative Enterprises

Abstract

In the context of global economic digital transformation and technological innovation, the application of AI Technology has a profound impact on corporate innovation and development. Existing research has primarily focused on the direct effect of AI Technology on Corporate Innovation Performance, while there is limited exploration of its interaction with organizational learning mechanisms. Based on the Dynamic Capabilities Theory, this study constructs a framework of “Technology—Individual Learning Capability—Team Learning Capability—Innovation Performance”, analyzing how AI Technology enhances learning capabilities to drive improvements in innovation performance and explores the moderating role of Organizational Learning Capability. Through empirical analysis of data from Specialized, Refined, Unique, and Innovative Enterprises in China, the study finds that AI Technology significantly enhances Corporate Innovation Performance, with Organizational Learning Capability playing a critical moderating role. Additionally, heterogeneity analysis indicates that factors such as production factors, industry characteristics, and firm size significantly influence the effectiveness of AI Technology in enhancing innovation performance. This research reveals the pathway through which AI Technology optimizes organizational learning mechanisms to improve innovation performance, offering both theoretical support and practical guidance for corporate strategic decision-making.

1. Introduction

In recent years, AI Technology has developed rapidly and has become a key driver of corporate innovation and industrial upgrading. Globally, AI Technology is recognized as a core pillar for enhancing economic competitiveness and achieving strategic breakthroughs [1]. Various countries have implemented national strategies to promote AI innovation; for example, the United States has strengthened funding support for AI research and its application through the “National Artificial Intelligence Research and Development Strategic Plan”, Japan has implemented the “Comprehensive Innovation Strategy 2024” to maintain its leading position in AI Technology, and Germany has reinforced AI research and development through the “Digital Strategy 2025”, focusing on integrating and applying Industry 4.0 technologies. At the enterprise application level, Tesla serves as a typical example of revolutionary innovation enabled by AI Technology, particularly in the fields of autonomous driving and intelligent manufacturing, significantly enhancing production efficiency and market competitiveness. Within this global context, Specialized, Refined, Unique, and Innovative Enterprises (SRUIEs) in China provide a distinct and significant example of AI adoption. These enterprises are characterized by their focus on specialized sectors, with unique operational strategies that combine niche market demands with cutting-edge technologies. SRUIEs are agile in their ability to adapt quickly to changing market conditions and technological trends, making them prime examples of firms leveraging AI for innovation. The integration of AI Technology within SRUIEs facilitates improvements in various operational processes, such as decision-making, data management, and the development of new products and services. AI plays a key role in enhancing the accuracy of predictions, personalizing customer interactions, and optimizing supply chain management, which in turn drives the firms’ innovation capacities. Organizational learning capability (Olc) is another crucial factor in the success of SRUIEs. Olc refers to the ability of an organization to continually learn, adapt, and implement new knowledge to improve its operations. Within SRUIEs, Olc is manifested through continuous employee training, knowledge-sharing mechanisms, and cross-functional collaboration. These learning processes enable the firms to integrate AI technologies more effectively, fostering an environment conducive to both incremental and breakthrough innovations. The dynamic capabilities framework further supports this relationship, emphasizing that organizational learning is essential for firms to respond to technological changes and sustain competitive advantage in rapidly evolving markets. Corporate innovation performance, in the context of SRUIEs, can be measured using several indicators (see

Table 1. Corporate innovation performance indicators for SRUIEs.

Indicator	Description	Method of Measurement
Innovation Rate	Reflects the number of new products or services launched in a specific period.	Count of new product/service introductions.
R&D Investment	Financial commitment to research and development activities.	Total R&D expenditure within a fiscal year.
Market Share Growth	Success of innovation in capturing market segments.	Percentage increase in market share.
Patent Applications	Traditional measure of technological innovation.	Number of patent applications filed.
New Business Models	Innovation in business strategies or models.	Identification of novel business models.
Customer Satisfaction	The impact of innovation on customer experience and satisfaction.	Surveys and customer feedback analysis.

). These include the innovation rate, which reflects the number of new products or services launched within a certain period, and R&D investment, which indicates the financial commitment towards research and development activities. Other important indicators include market share growth, which signifies the success of innovation in capturing market segments, and patent applications, which serve as an established measure of technological innovation. While patents provide a traditional way of measuring innovation, they are often complemented by other qualitative metrics to offer a more holistic view of a firm’s innovative capabilities. The combination of AI Technology, Organizational Learning Capability, and Corporate Innovation Performance in SRUIEs highlights the pivotal role that technology plays in enhancing business competitiveness. As these firms continue to integrate AI into their operations, they demonstrate the significant impact of AI on innovation, particularly in terms of improving efficiency, product development, and market competitiveness. This section not only deepens the understanding of how SRUIEs in China utilize AI to drive innovation but also provides a broader perspective on the integration of AI in organizational strategies, offering valuable insights into the intersection between dynamic capabilities and technological innovation.

Existing studies exploring the impact mechanisms of AI Technology on Corporate Innovation Performance can be categorized into three main areas: information processing capability, dynamic adaptability, and collaborative innovation capability. These categories were developed based on the distinct ways in which AI Technology influences organizational learning and innovation. Each area provides valuable insights into how AI enhances organizational capabilities, which ultimately drives innovation outcomes. Regarding information processing capability, Cui et al. (2022) [2] examined the role of information processing capability in precision agriculture technology, finding that the quality of technology adoption is influenced by information processing control and network paths. Huber et al. (2022) [3] proposed an information systems capability framework that helps manufacturing industries adapt to information technology transformations. Lu et al. (2023) [4] explored the role of supply chain information processing capability in enhancing supply chain resilience, highlighting that improving this capability significantly enhances risk resistance, with supply chain governance playing a mediating role in the process. Yu et al. (2022) [5] found that enhancing the openness of technological innovation improves information processing capability, thereby strengthening supply chain resilience and improving corporate performance. These studies are representative of how AI Technology integrates into organizations to enhance their information processing systems. Regarding dynamic adaptability, Chatterjee et al. (2022) [6] found that the dynamic capabilities of frontline employees significantly enhance adaptability, which in turn drives organizational performance improvement. Ghahramani et al. (2023) [7] emphasized that organizational absorptive capacity, by improving adaptability, contributes to continuous improvements in information security management, especially in environments with high competitive pressure. Xu et al. (2024) [8] indicated that improvements in technological and institutional adaptability drive human capital accumulation, fostering economic development, particularly in technology-intensive and high internet penetration areas. Roh et al. (2022) [9] found that organizational design capabilities and employee communication skills enhance supply chain adaptability and responsiveness. These studies highlight the importance of dynamic adaptability in leveraging AI Technology for organizational development. Finally, collaborative innovation capability is another critical area of research. Xie et al. (2023) [10] conducted a meta-analysis and found a significant positive correlation between supply chain collaborative innovation and innovation performance, with collaboration in the industry–university–research institute model being particularly prominent. Wei et al. (2023) [11] discovered that three relational strategies—super modular complementarity, unique complementarity, and consistency dependence—within platform-based innovation ecosystems promote collaborative innovation, with AI Technology enhancing cooperation and innovation performance within platform enterprises. Dolmans et al. (2023) [12] analyzed collaborative innovation in smart cities from a dynamic perspective and pointed out that AI Technology can effectively facilitate cross-departmental collaboration, particularly in complex environments such as smart cities, where overcoming innovation barriers is a key aspect of AI applications. These studies underscore the role of AI Technology in driving collaborative innovation and enhancing organizational performance in complex systems. The studies categorized in this review are representative as they include empirical research across diverse industries and perspectives. Notably, the focus on Specialized, Refined, Unique, and Innovative Enterprises (SRUIEs) in China offers unique insights into the application of AI Technology within SMEs. SRUIEs represent a critical subset of SMEs with significant innovation potential, despite facing challenges such as resource constraints and limited technological capabilities. By focusing on these enterprises, this research aims to explore how AI Technology can enhance Organizational Learning Capability and, by extension, Corporate Innovation Performance. The findings from this research offer valuable guidance for SMEs globally, providing insights into how they can adapt AI Technology to overcome their inherent challenges.

In terms of research methods, existing studies typically employ regression analysis, Structural Equation Modeling (SEM), and case studies to explore the impact of AI Technology on Corporate Innovation Performance. First, in regression analysis, Zhang et al. (2023) [13] used multiple linear regression to investigate the impact of AI Technology on Corporate Innovation Performance, finding that AI Technology significantly enhanced the technological innovation capabilities of enterprises, particularly in areas such as research and development efficiency, product design, and market adaptability. Jiang et al. (2023) [14] explored the interactive effects of AI and service transformation in the manufacturing industry, revealing that AI played a key role in driving technological innovation performance in the sector. Second, studies using Structural Equation Modeling have also highlighted the significant relationships between innovation and performance. Espasandín-Bustelo et al. (2023) [15] applied SEM to explore the relationship between socioeconomic enterprise innovation and performance, showing that innovation outcomes are influenced by managerial capabilities and customer recognition of the legitimacy of innovations. The study confirmed a significant relationship between innovation performance and economic outcomes. Merín-Rodríguez et al. (2024) [16] employed PLS-SEM to analyze the impact of digital transformation on the performance of innovative SMEs, discovering that business model innovation partially mediates the relationship between digital transformation and corporate performance. Finally, case study methods have provided valuable insights for research. Lu et al. (2024) [17] conducted a multi-case study to examine the impact of ambidextrous innovation on the financing performance of SMEs, finding that both exploratory and exploitative innovations improve financing performance, with the breadth and depth of digital technologies moderating their effects. Marín et al. (2023) [18] employed case study methods to explore the effects of digital transformation and cross-value chain collaboration on SMEs’ innovation, showing that digital transformation and cross-chain collaboration significantly promoted technological innovation and patent applications among SMEs.

From a theoretical perspective, the theories of technological affordance and sociotechnical systems provide two core frameworks for examining the impact of AI Technology on Corporate Innovation Performance. The theory of technological affordance emphasizes that the availability of technological resources plays a crucial role in driving innovation within enterprises. Sun et al. (2024) [19] argue that digital transformation in enterprises, through information homogenization, reorganization, and conversion, facilitates green innovation and low-carbon energy breakthroughs. Ballerini et al. (2023) [20] examine the role of digital platforms in the e-commerce performance of SMEs in the manufacturing industry, finding that digital platforms, through affordances such as consumer knowledge generation, internationalization, and customer diversification, promote digital transformation and enhance corporate performance. From the sociotechnical systems theory perspective, Gillani et al. (2024) [21] reveal the repetitive patterns and typical characteristics exhibited by different enterprises during their digital transformation, highlighting that resource allocation and the enhancement of dynamic capabilities are key factors for successful transformation. The effective utilization of technological affordance provides strong support for corporate innovation. Marinakis et al. (2024) [22] emphasize that, during digital transformation, enterprises must not only rely on the introduction and application of external technologies but also focus on the reconfiguration and utilization of internal resources to gain an innovation advantage at various stages of the transformation.

Despite the widespread acknowledgment of the positive impact of AI Technology on Corporate Innovation Performance, several important gaps remain in the existing literature. First, most studies have focused on the direct effects of AI Technology on information processing and decision-making efficiency, without sufficiently exploring how AI influences Individual Learning Capability and Team Learning Capability through the optimization of organizational learning mechanisms. This study fills this gap by investigating how AI enhances these learning capabilities, which in turn strengthens the overall innovation capacity. The novel contribution of this research lies in the focus on organizational learning mechanisms as an intermediary between AI Technology and innovation performance. This approach offers a new perspective on how AI contributes to innovation, expanding the current understanding of AI’s role beyond mere efficiency improvements. Second, while existing research often relies on quantitative methods, such as regression analysis, to examine the relationship between AI and innovation performance, these studies fail to provide a complete understanding of the underlying mechanisms. This study innovates by integrating both quantitative and qualitative approaches, allowing for a more holistic examination of the interaction between AI Technology and organizational factors such as structure, culture, and learning mechanisms. The inclusion of qualitative data, such as case studies and interviews, enriches the analysis and provides a deeper understanding of how these factors influence AI’s effectiveness in driving innovation. This mixed-methods approach represents a novel contribution to the literature, offering a comprehensive framework for analyzing AI’s impact on organizational innovation. Third, while much of the existing literature has focused on the alignment between technology and organizational structure, there is limited attention to how firms adapt to technological changes through capability adjustments and resource reorganization. The innovation of this study lies in its application of the Dynamic Capabilities Theory to investigate how AI Technology facilitates organizational adaptation. This study examines how firms reconfigure their capabilities and reorganize resources in response to technological advancements, thereby transforming their innovation outcomes. By applying this theory to the context of AI, the study offers new insights into how organizations leverage dynamic capabilities to manage technological disruptions and enhance innovation. This theoretical application is a significant departure from previous research that primarily examines the alignment of technology and structure without considering the dynamic processes through which firms adapt. Finally, this study contributes to the literature by focusing on Specialized, Refined, Unique, and Innovative Enterprises (SRUIEs) in China, a category of small and medium-sized enterprises (SMEs) that has received limited attention in previous research. While much of the existing literature on AI adoption and innovation focuses on large corporations or firms in developed economies, SRUIEs in China present a unique opportunity to explore how resource-constrained SMEs leverage AI to improve innovation. This study provides valuable insights into how such SMEs, which face different challenges and opportunities compared to large enterprises, can utilize AI to drive innovation and compete in global markets. The focus on SRUIEs represents an innovative aspect of the study, offering context-specific contributions to the understanding of AI adoption in emerging economies.

Grounded in the Dynamic Capabilities Theory, this study aims to empirically investigate how AI Technology enhances both Individual and Team Learning Capability, ultimately improving Corporate Innovation Performance. Using data from Specialized, Refined, Unique, and Innovative Enterprises (SRUIEs) in China, the study seeks to answer the following three key questions: First, does AI Technology significantly impact Corporate Innovation Performance? If so, is the impact positive or negative? Second, what is the specific mechanism through which AI Technology influences Corporate Innovation Performance? Third, do factors such as enterprise type, industry characteristics, and enterprise size still have a significant effect on the application of AI Technology and its impact on Corporate Innovation Performance?

2. Theoretical Foundation and Basic Hypotheses

2.1. Theoretical Foundation

The Dynamic Capabilities Theory [23] posits that firms must possess the flexibility to adjust their resource allocation in response to rapidly changing external environments in order to maintain sustained competitive advantage and innovation performance. This theory emphasizes that, in addition to accumulating and utilizing existing resources, firms must adapt to environmental changes by reorganizing resources to address challenges related to market demand and technological innovation. Eisenhardt and Martin (2000) [24] further argued that dynamic capabilities not only encompass resource integration but also involve the ability to perceive changes in the external environment, formulate response strategies, and restructure mechanisms, highlighting the synergistic role of technology and learning capabilities in optimizing resource allocation and enhancing innovation capacity. Zollo and Winter (2002) [25] pointed out that Organizational Learning Capability is at the core of dynamic capabilities. By integrating Individual Learning Capability and Team Learning Capability, firms can rapidly absorb external knowledge, optimize internal resource allocation, and drive continuous innovation. In the context of digital transformation, AI Technology offers a new perspective on enhancing dynamic capabilities. Teece (2014) [26] argued that, in the digital era, dynamic capabilities rely not only on traditional resource integration but also on leveraging emerging technologies to improve learning efficiency and innovation performance. AI Technology enhances Individual Learning Capability through efficient data analysis and real-time feedback while optimizing Team Learning Capability through improved collaboration and knowledge sharing. This, in turn, further drives organizational innovation capacity [27]. In conclusion, Organizational Learning Capability plays a key moderating role between technology application and learning capabilities, converting Individual Learning Capability into a team advantage. This facilitates the flow and sharing of knowledge, allowing the full potential of AI Technology to be harnessed within teams and enhancing the firm’s adaptability and innovation performance in complex environments.

Based on this, this study proposes the “Technology—Individual Learning Capability—Team Learning Capability—Performance” analytical framework. Specifically, AI Technology, as the core variable, directly impacts Corporate Innovation Performance by enhancing Individual Learning Capability and optimizing team collaboration. Individual Learning Capability promotes knowledge accumulation and transformation, providing foundational support for Team Learning Capability. Team Learning Capability further integrates individual knowledge, fosters collaboration, and drives innovation, thus improving organizational performance. Organizational Learning Capability, as a moderating variable, forms a synergistic effect between technology application and learning capabilities, thereby enhancing the driving effect of AI Technology on innovation performance. This framework reveals the interactive relationship between technology, learning capabilities, and performance, elucidating how firms can achieve competitive advantage through dynamic capabilities.

2.2. Research Hypotheses

Drawing from the Dynamic Capabilities Theory, a firm’s innovation performance hinges on its ability to rapidly sense, integrate, and reconfigure resources in response to dynamic market environments, ensuring the maintenance of a competitive advantage. In this context, AI Technology is viewed as a significant enabler of innovation, fostering faster adaptation and enhancing decision-making efficiency, which in turn catalyzes innovation [28]. The effective implementation of AI not only optimizes resource allocation but also boosts the efficiency of technological and product innovations, thus significantly enhancing overall innovation capabilities.

In the specific setting of China, Specialized, Refined, Unique, and Innovative Enterprises (SRUIEs) face unique challenges from both technological evolution and market competition. AI Technology is increasingly applied in R&D, production processes, and market analysis, improving R&D efficiency and boosting the innovation capacity of these enterprises [29]. With the help of advanced AI systems such as smart production tools, data analytics, and decision support mechanisms, firms can more accurately forecast market demand, adjust product designs swiftly, accelerate technological innovations, and achieve improvements in both productivity and innovation outcomes.

Moreover, SRUIEs with robust technological adoption and Organizational Learning Capability are more adept at absorbing and applying AI, further enhancing their innovation potential [30]. These firms, backed by strong dynamic capabilities, integrate AI into their innovation systems more effectively, promoting deepened innovation through continuous optimization, and thereby standing out in competitive markets with superior innovation performance. This suggests the following refined hypothesis:

Hypothesis 1. 

AI Technology positively impacts the Corporate Innovation Performance of Specialized, Refined, Unique, and Innovative Enterprises, with Organizational Learning Capability playing a significant moderating role in enhancing innovation outcomes.

According to the Dynamic Capabilities Theory, Corporate Innovation Performance is closely related to Organizational Learning Capability. This theory emphasizes that firms must possess the ability to rapidly sense and respond to new technologies, with Organizational Learning Capability serving as a key enabler of this ability. Specifically, Organizational Learning Capability not only helps firms enhance their technological application capabilities through experiential learning but also facilitates the absorption, integration, and innovation of external information, thereby enabling firms to quickly adapt to market changes and enhance their innovation performance. AI Technology, through big data analysis and intelligent decision-making optimization, enhances a firm’s innovation capacity and performance [31].

However, the effective application of AI Technology depends not only on the technology itself but also on the firm’s Organizational Learning Capability. Whether Specialized, Refined, Unique, and Innovative Enterprises can effectively integrate AI Technology into their innovation systems and continuously optimize it through organizational learning determines the extent of their innovation performance. Even if a firm possesses high technological adoption capabilities, the full potential of AI Technology cannot be realized without strong Organizational Learning Capability. In contrast, firms with robust Organizational Learning Capability can more rapidly absorb and optimize AI Technology, thereby improving innovation performance.

Moreover, Organizational Learning Capability enables firms to extract valuable insights from technological innovations, driving technological iteration and innovation through continuous reflection and adjustment [32]. Therefore, AI Technology, underpinned by Organizational Learning Capability, can more effectively enhance Corporate Innovation Performance. Based on the above analysis, the following hypotheses can be proposed:

Hypothesis 2. 

Organizational Learning Capability positively moderates the impact of AI Technology on the Corporate Innovation Performance of Specialized, Refined, Unique, and Innovative Enterprises.

In light of this, the theoretical model of this study is illustrated in Figure 1.

3. Research Design

3.1. Data Sources

This study, based on Hossain et al. (2024) [33] regarding the digitalization of strategic innovation in small and medium-sized enterprises (SMEs, selects listed Specialized, Refined, Unique, and Innovative Enterprises (SRUIEs) in China from 2013 to 2023 as the initial sample to ensure the accuracy and rigor of the data. The study follows the guidelines outlined in the “Notice on Further Supporting the High-Quality Development of Specialized, Refined, Unique, and Innovative SMEs” issued by the Ministry of Finance and the Ministry of Industry and Information Technology in 2024 in conjunction with Ramdani et al. (2023) [34], while considering the availability of data to ensure both the broadness and representativeness of the data sources.

In terms of data screening, this study follows the criteria set by Le and Behl (2023) [35], selecting SRUIEs that were continuously listed and published annual reports between 2013 and 2023. Enterprises with missing data or significant financial anomalies were excluded. Ultimately, the study obtained balanced panel data from 1657 companies, totaling 18,227 valid observations. To ensure data quality, all continuous variables were trimmed at the 1% level to mitigate the influence of outliers.

The data sources primarily include the Guotai An Database, annual reports of listed companies, the China Statistical Yearbook, and the Wind Financial Database, which ensure the reliability, comprehensiveness, and timeliness of the data.

3.2. Variable Measurement

3.2.1. Core Explanatory Variable

The core explanatory variable in this study is AI Technology, measured by the breadth and depth of its application in production, R&D, and management. Unlike most studies that rely on survey data [36,37], this research does not adopt this method primarily due to the susceptibility of surveys to subjective bias and cognitive differences. Additionally, surveys are limited in providing large-scale, long-term dynamic data and are less precise in distinguishing between firm performance in areas such as technological R&D and patent applications. In contrast, patent data offer greater objectivity and systematic reliability, accurately reflecting a firm’s technological accumulation, innovation capabilities, and R&D investment in the AI field. As a measurement standard, patent data provide reliable evidence to uncover a firm’s technological strategy and innovation practices [38].

This study quantifies patent data to explore the application of AI Technology in corporate management and its impact on innovation capability. First, a literature review defines and classifies AI Technology, identifying key terms related to technological innovation, such as machine learning, big data analytics, and cloud computing [39,40]. To overcome the limitations of manually selected keywords, natural language processing techniques are employed to expand the keyword database, resulting in 150 keywords, which are categorized into basic technologies, application technologies, and management technologies. Unlike studies that rely on enterprise surveys or publicly available databases [41,42], this research utilizes patent texts of listed companies as a corpus, applying optical character recognition (OCR) technology to extract technical abstracts and analyze the firms’ technological layout in AI (see

Table 2. Keywords for AI Technology.

Variable Name	Seed Words	Expanded Seed Words	References
AI Technology	Artificial Intelligence, AI, Fundamental Algorithms, Neural Networks, Data Mining, Natural Language Processing, Deep Learning, Machine Learning, Reinforcement Learning, Supervised Learning	Deep Learning Models, Machine Learning and AI, Deep Reinforcement Learning, Machine Learning Technology, Deep Learning, Unsupervised Learning, Natural Language Processing, Information Mining, Deep Learning Networks, Neural Network Training, Semi-supervised Learning, Natural Language Generation, Large-scale Machine Learning, Unsupervised Learning, Neural Networks, Future AI, Supervised Learning, Artificial Neural Networks, Big Data Mining, Supervised Learning, AI Robots, Machine Learning Models, Pattern Recognition, Natural Language Technologies, Machine Intelligence, Convolutional Neural Networks, NLP Technologies, Data Visualization, Data Analysis and Mining, Big Data Analysis and Mining, AI Technology, Computer Vision, Neural Network Models, Text Mining, Applied Machine Learning, AI and Machine Learning, Deep Neural Networks, Deep Self-learning, Machine Learning Algorithms, AI Deep Learning, Neural Network Structures, Machine Learning Domains, Deep Learning Technologies, Deep Learning Systems, Natural Language Processing Technologies, Deep Learning Algorithms, Data Mining and Analysis, AI AI, Neural Network Systems, Automated Machine Learning, Image Recognition, Natural Language Understanding, NLP, Natural Language Processing, Neural Network Algorithms, Machine Learning and Deep Learning, Internet of Things, Data Science, Deep Learning Neural Networks, Neural Networks, Recurrent Neural Networks, AI and Robotics, Data Modeling, Deep Neural Networks, AI and Machine Learning, Generative Models, Machine Learning, Semantic Understanding, AI and Deep Learning, Knowledge Graphs, Convolutional Networks, Speech Processing, Transfer Learning, Data Analysis, Learning Algorithms, Language Processing	[43,44,45]

). Finally, word segmentation and Z-score standardization methods are used to analyze word frequency data, ensuring the robustness and reliability of the measurement results.

3.2.2. Dependent Variable

The dependent variable, Corporate Innovation Performance, measures a firm’s comprehensive performance across multiple innovation domains, including products, technologies, and services [46]. Currently, three primary methods are used to assess innovation performance: first, measuring technological innovation through the number of patents [47]; second, measuring innovation input via the ratio of R&D expenditure to revenue [48]; and third, using a multidimensional indicator system to comprehensively assess product and service innovation [49]. Given its comprehensiveness, this study adopts the third method. Following Li and Wei’s (2024) framework for measuring digital transformation [50], a three-tier indicator system is constructed (see

Table 3. Three-tiered indicator system for Corporate Innovation Performance.

First-Level Indicator	Second-Level Indicator	Third-Level Indicator	Reference
Corporate Innovation Performance	Product Innovation Performance	Ratio of new product sales to total revenue	[51]
		Natural logarithm of the sum of utility model patents and design patents, plus one	[52]
	Service Innovation Performance	Ratio of R&D expenditure to total revenue	[53]
		Price-to-Earnings (P/E) ratio	[54]

), encompassing product innovation performance and service innovation performance, while accounting for both R&D investment and market performance. Specifically, product innovation performance is measured by the proportion of new product sales to total revenue and the number of patents (the natural logarithm of the sum of utility model patents, design patents, and one), where the former reflects market competitiveness and the latter integrates technological and design innovation. Service innovation performance is measured by the ratio of R&D expenditure to revenue and the price-to-earnings ratio. The former reflects the allocation of innovation resources, while the latter gauges market expectations of the firm’s innovation potential. Finally, the entropy method is employed to assign weights to each indicator and aggregate them, ensuring the scientific and comprehensive nature of the evaluation. This approach overcomes the limitations of single-dimensional indicators, providing a more precise assessment of innovation performance.

The specific steps for calculating the entropy weighting method are as follows:

First, calculate the entropy value. The entropy value is a measure of the uncertainty of a random variable. The formula for calculating the entropy value is as follows:

$$E_j=-k\ast {\displaystyle \sum [p(r_j)}\ast \ln (p(r_j))]$$

(1)

In this context, $p(r_j)={\displaystyle \frac{r_{ij}}{{\displaystyle \sum r_{ij}}}}$ represents the ratio of the $j$ object in terms of the indicator, which is its normalized value divided by the sum of the normalized evaluation values for all evaluation objects on that indicator. $E_j$ denotes the entropy value for the indicator, used to measure the uncertainty of the object with respect to that indicator. $k$ is the adjustment coefficient for the entropy calculation.

Second, the coefficient of variation is computed to indicate the degree of variation of an object on a particular indicator. The formula for this calculation is as follows:

$$d_j=1-E_j$$

(2)

Third, the calculation of weights is based on the coefficient of variation to determine the weight of each indicator. Indicators with higher weights have a greater influence on the evaluation. The formula for this calculation is as follows:

$$w_i={\displaystyle \frac{d_j}{{\displaystyle \sum d_j}}}$$

(3)

Fourth, the calculation of the composite score involves multiplying the value of each object for each indicator by the corresponding weight of that indicator and then summing the results to obtain the composite score for each object. The formula for this calculation is as follows:

$$s_j={\displaystyle \sum (w_i\ast x_{ij}})$$

(4)

where $s_j$ represents the composite score of the $j$ object; $w_i$ denotes the weight of the $i$ indicator; and $x_{ij}$ is the value of the $j$ object on the $i$ indicator.

Finally, based on the weights, the raw data are transformed into composite scores, thereby constructing the comprehensive measurement indicator data for Corporate Innovation Performance.

3.2.3. Moderator Variable

Organizational Learning Capability refers to a firm’s ability to acquire, understand, share, and apply knowledge in dynamic environments, directly influencing its innovation capacity and ability to respond to competitive pressures. Unlike simple financial indicators, Organizational Learning Capability effectively measures a firm’s sustainability in learning and innovation [30]. Therefore, based on the existing literature, this study constructs a three-tier indicator system (see

Table 4. Three-level indicator system of Organizational Learning Capability.

First-Level Indicator	Second-Level Indicator	Third-Level Indicator	Reference
Organizational Learning Capability	Individual Learning Capability	The natural logarithm of the number of invention patents granted, plus one	[55]
		Proportion of employees with a master’s degree or higher	[56]
	Team Learning Capability	Sales revenue ratio	[57]
		Growth rate of operating revenue	[58]

) and utilizes the entropy method to generate an Organizational Learning Capability index, providing a comprehensive reflection of a firm’s Organizational Learning Capability.

Drawing upon the research of Greenan et al. (2024) [59] and Dohse et al. (2023) [55], Individual Learning Capability is measured by the number of granted invention patents and the proportion of employees with postgraduate or higher education. The number of granted patents is smoothed using the natural logarithm of the patent count plus one to reduce heteroscedasticity, while the education level is assessed based on the proportion of employees with postgraduate or higher degrees [56]. Team Learning Capability is represented by two indicators: sales return rate and operating income growth rate. The former reflects the team’s innovative achievements in market promotion, while the latter reveals the team’s ability to drive innovation and expand the market.

This study further refines the first-level indicators of Individual Learning Capability and Team Learning Capability and clarifies the second-level indicators (e.g., invention patent grants, education level, sales return rate, operating income growth rate, etc.) [60]. Through a quantitative approach, a third-level indicator system is constructed, ensuring that each indicator accurately reflects the firm’s capabilities in organizational learning and providing theoretical support for the subsequent empirical analysis.

3.2.4. Control Variables

Drawing on the approaches of Long et al. (2024) [61], Ma et al. (2024) [62], Dey et al. (2024) [63], and Alouane et al. (2022) [64], this study controls for several factors that may influence Corporate Innovation Performance during the analysis process. Specifically, the control variables included are as follows: ① Firm Size (Size), measured by total assets, reflecting the firm’s financial strength and market position; ② Firm Age (Age), calculated as the difference between the current year and the firm’s listing year, reflecting the firm’s maturity and accumulated experience; ③ Firm Growth (Gro), measured by the growth rate of total assets, indicating the firm’s development potential and innovative capacity; ④ Current Assets Ratio (Flu), calculated as the ratio of current assets to total assets, reflecting the firm’s short-term solvency and financial health; ⑤ Return on Equity (ROE), measured as the ratio of net profit to shareholder equity, assessing the firm’s profitability and capital utilization efficiency; ⑥ Ownership Concentration (Owner), measured by the proportion of shares held by the largest shareholder, influencing the firm’s decision-making efficiency and governance structure. The inclusion of these control variables helps to effectively eliminate the interference of other potential factors, thereby enhancing the explanatory power and predictive accuracy of the research model (see

Table 5. Control variables and their measurement methods.

Variable Name	Abbreviation	Measurement Method	Reference
Firm Size	Size	Total assets of the firm	[61]
Firm Age	Age	Current year minus the year of the firm’s listing	[61]
Firm Growth	Gro	Growth rate of total assets	[62]
Current Asset Ratio	Flu	Current assets/Total assets	[63]
Return on Equity (ROE)	ROE	Net profit/Shareholders’ equity balance	[63]
Ownership Concentration	Owner	Proportion of shares held by the largest shareholder	[64]

).

3.3. Model Specification

This study primarily examines the impact of AI Technology on Corporate Innovation Performance. Using firm-level data, panel data are constructed in accordance with relevant studies and the context of this research [65]. A fixed-effects model is employed for regression estimation, and the baseline regression model is specified as follows:

$$Ol{c}_{i,t}={\beta }_0+{\beta }_{1}A{I}_{i,t}+{\beta }_3{\displaystyle \sum Contro{l}_{i,t}}+{\mu }_i+{\nu }_t+u_{i,t}$$

(5)

In this context, the subscripts $i$ and $t$ represent the firm and the year, respectively. The dependent variable, $Ol{c}_{i,t}$, denotes the innovation performance of firm $i$. The core explanatory variable, AI, represents the level of $A{I}_{i,t}$ Technology adoption by firm $i$ in year $t$, with a higher value indicating a greater degree of AI Technology application by the firm. $Contro{l}_{i,t}$ denotes the set of control variables. ${\mu }_i$ and ${\nu }_t$ represent the industry and year fixed effects, respectively. $u_{i,t}$ refers to the error term.

4. Empirical Results

4.1. Descriptive Statistics and Baseline Regression

4.1.1. Descriptive Statistics

To gain a deeper understanding of the variables under investigation and their distribution characteristics, this study first conducts a descriptive statistical analysis.

Table 6. Descriptive statistics.

VarName	Obs	Mean	SD	Min	Max
Ie	18,227	59.04	37.49	26.01	287.20
AI	18,227	1.83	5.89	0.00	40.00
Size	18,227	23.7	22.9	5.16	160.00
Age	18,227	16.40	4.72	7.00	30.00
Gro	18,227	0.19	0.22	−0.19	1.24
Flu	18,227	0.64	0.14	0.25	0.95
ROE	18,227	0.08	0.07	−0.32	0.27
Owner	18,227	29.62	13.23	0.47	63.33

presents the statistical descriptions of the key variables, including their observed values, means, standard deviations, minimum values, and maximum values. Regarding Corporate Innovation Performance (Olc), the mean value is 59.04, with a standard deviation of 37.49. This indicates a significant variance in Corporate Innovation Performance, reflecting the disparity in the innovative activities of different firms. Furthermore, the minimum value of Corporate Innovation Performance is 26.01, while the maximum value reaches 287.20, highlighting that some firms excel in innovation, while others may face challenges in enhancing their innovative capabilities. For AI Technology, the mean value is 1.83, with a standard deviation of 5.89. The large standard deviation suggests that, within the sample, some firms are at a high level of AI Technology adoption, whereas others have minimal or no investment or application in AI Technology. This implies that the diffusion of AI Technology varies significantly across firms, and its potential impact warrants further investigation. Overall, the data for the key variables fall within reasonable ranges, providing a solid foundation for subsequent analysis.

Regarding the control variables, the mean value of firm size (Size) is 23.7, with a standard deviation of 22.9, indicating significant differences in firm size across the sample. The smallest firm has an asset size of 5.16, while the largest firm’s asset size reaches 160.00, suggesting the presence of economies of scale within the sample. The mean value of firm age (Age) is 16.40, with a standard deviation of 4.72, reflecting a certain concentration in the firms’ life cycles. Most firms have been established for a relatively long period and have undergone market validation. The mean value of firm growth (Gro) is 0.19, with a standard deviation of 0.22, indicating that some firms exhibit strong growth performance, while others face challenges in terms of growth stagnation. The mean value of the current asset ratio (Flu) is 0.64, suggesting that most firms possess sound liquidity management capabilities, which help them meet short-term financial needs. The mean value of the Return on Equity (ROE) is 0.08, indicating generally favorable capital utilization efficiency across firms, although some firms show poor performance in terms of profitability. Finally, the mean value of ownership concentration (Owner) is 29.62, with a standard deviation of 13.23, highlighting that variations in ownership structure across firms may impact decision-making efficiency and innovation motivation. All these key control variables fall within reasonable ranges and provide essential foundational information for subsequent empirical analysis.

4.1.2. Benchmark Regression

The parameter estimation results of the benchmark model are presented in

Table 7. Baseline regression results.

	I	II	III
	Model 1	Model 2	Model 3
AI	0.3425 ***	0.2127 ***	0.2755 ***
	(0.0610)	(0.0710)	(0.0748)
Size		−0.0000 ***	−0.0000 ***
		(0.0000)	(0.0000)
Age		−0.1866 ***	−0.6920 ***
		(0.0641)	(0.0850)
Gro		16.8507 ***	3.6788 **
		(1.5459)	(1.5379)
Flu		−4.4220 **	−2.7543
		(2.2308)	(2.4471)
ROE		−1.2 × 102 ***	−62.9141 ***
		(4.9406)	(5.2124)
Owner		0.0602 **	0.0998 ***
		(0.0280)	(0.0298)
_cons	58.4149 ***	72.2362 ***	54.1809 ***
	(0.2800)	(2.1530)	(5.1426)
Control Variables	NO	YES	YES
Year Fixed Effects	NO	NO	YES
Industry Fixed Effects	NO	NO	YES
N	18,227	18,227	18,227
R2	0.003	0.07	0.348

** p < 0.05, *** p < 0.01.

. Column I shows the estimates without the inclusion of control variables, Column II displays the estimates with control variables included, and Column III shows the estimates after controlling for year and industry effects.

Firstly, for the core explanatory variable, AI Technology, a significant positive correlation is observed in all three models. Specifically, the coefficient in Model 1 is 0.3425, in Model 2 it is 0.2127, and in Model 3 it is 0.2755. This trend suggests that with the introduction of control variables, the effect of AI on Corporate Innovation Performance slightly weakens but remains statistically significant. This indicates that although the impact of AI Technology on Corporate Innovation Performance varies across different model specifications, it consistently exerts a positive influence, thereby validating Hypothesis 1. According to Wu et al. (2024) [66], technological innovation is a key driver of sustained corporate growth. Based on this finding, the present study further infers that in the innovation processes of Specialized, Refined, Unique, and Innovative Enterprises (SRUIEs), there exists a positive relationship between technological application and innovation performance, implying that technological innovation contributes to enhancing corporate performance, thus influencing market competitiveness.

Secondly, regarding the control variables, firm size is significantly negatively correlated with innovation performance in Models 2 and 3, suggesting that larger firms may face reduced innovation efficiency due to complex organizational structures and decision-making processes. Firm age also shows a negative correlation, indicating that older firms may lack the flexibility to adopt new technologies. High-growth firms demonstrate significantly higher innovation performance, highlighting their greater innovation potential and market adaptability. As for other variables such as the liquidity ratio (Flu), Return on Equity (ROE), and ownership concentration (Owner), the data reveal a complex relationship with Corporate Innovation Performance, with varying effects across different models. The associations between these variables and innovation strategies and performance provide further nuanced insights for this study.

Additionally, while the robustness tests address factors such as the replacement of the dependent variable and exclusion of COVID-19 data, it is acknowledged that other potential confounders, such as economic conditions, regional variations, industry-specific shocks, and government policy interventions, could also influence the results. These factors have not been directly incorporated into the current analysis but should be considered in future research to further enhance the robustness of the findings.

4.2. Robustness Check and Endogeneity Analysis

4.2.1. Robustness Check

To ensure the accuracy and reliability of the research findings, several robustness tests were conducted in this study, addressing measurement errors, the exclusion of special factors, and omitted variable bias [67,68,69]. First, to address measurement errors, the dependent variable in Model 4 was replaced with overall corporate innovation quality. The results show that the positive impact of AI Technology on innovation quality remains significant, with a coefficient of 0.0022, significant at the 1% level, thus confirming the robustness of the baseline regression results. See

Table 8. Robustness test.

	I	II	III
	Model 4	Model 5	Model 6
AI	0.0022 ***	1.0836 *	0.4331 *
	(0.0004)	(0.5572)	(0.2568)
Control Variables	YES	YES	YES
Year Fixed Effects	YES	YES	YES
Industry Fixed Effects	YES	NO	YES
N	18,227	18,227	18,227
R2	0.502	0.001	0.006

* p < 0.1, *** p < 0.01.

.

Second, to eliminate the impact of special factors, Model 5 excluded data from the COVID-19 pandemic in 2021. The results indicate that after exclusion, the coefficient of AI Technology is 1.0836, still significantly positive, demonstrating that the positive effect of AI Technology on Corporate Innovation Performance is not influenced by extreme data fluctuations during atypical periods (see Table 8).

Third, to control for the influence of omitted variables, Model 6 used a fixed-effects model. The results show that the coefficient of AI Technology is 0.4331, and after controlling for firm-specific characteristics, it still exerts a significant positive effect on innovation performance, further validating the robustness of the study’s conclusions (see Table 8).

In summary, through three robustness checks, the models in this study consistently exhibit strong robustness under different conditions, demonstrating the persistent and significant positive effect of AI Technology on Corporate Innovation Performance. This provides solid theoretical support for firms to increase investment in AI Technology and enhance their innovation capabilities.

4.2.2. Endogeneity Analysis

To address the issue of endogeneity, this study employs the instrumental variables (IV) method, following the approach of Tian et al. (2023) [70], using lagged values of AI Technology as instruments. Specifically, AI Technology data for one and two periods lagged are selected as instrumental variables, and the results of this test are presented in

Table 9. Instrumental variable regression results.

	I	II	III
	Model 7	Model 8	Model 9
AI	0.2755 ***
	(4.41)
L1AI		0.9610 ***
		(68.79)
L2AI			0.9889 ***
			(44.68)
Control Variables	YES	YES	YES
Year Fixed Effects	YES	YES	YES
Industry Fixed Effects	YES	YES	YES
N	14,623	14,623	14,623
F-Statistic for Joint Test of Instrumental Variables (p-value)		4731.62
		(0.0000)
Under-identification Test: Anderson Canon. LM (p-value)			355.05
			(0.0000)
Weak Instrument Test: Cragg-Donald Wald F Statistic			28,956.26
Stock–Yogo Weak ID Test Critical Values: 10% Maximal IV			16.38

Standard errors in parentheses. *** p < 0.01.

.

In the model, the second column uses lagged one-period AI Technology as an instrumental variable. The results indicate a coefficient of 0.9610, which is statistically significant at the 1% level, suggesting a strong correlation between the instrument and the dependent variable. Additionally, the joint test of the instrument yields an F-statistic of 4731.62, with a p-value close to 0.0000, further confirming the validity of the instrumental variable.

In the third column, lagged two-period AI Technology is used as the instrument, controlling for other firm characteristics such as size and age. The results show that the coefficient of AI Technology is 0.9889, significant at the 1% level. Furthermore, the Anderson Canon. LM statistic is 355.05, with a p-value near 0.0000, indicating that there is no issue of under-identification and supporting the “relevance” assumption. The Cragg–Donald Wald F-statistic is 28,956.26, well above the critical value of 16.38 for the 10% threshold in the Stock–Yogo Weak ID Test, thus confirming the “exogeneity” assumption of the instrumental variable.

In conclusion, through the endogeneity analysis using the instrumental variables method, this study confirms the robustness of the positive association between AI Technology and Corporate Innovation Performance, providing theoretical support for the moderating role of Organizational Learning Capability in the innovation process.

4.3. Heterogeneity Analysis

4.3.1. Heterogeneity Analysis Based on Production Factors

Considering the critical influence of the factor structure on innovation performance, this study classifies enterprises into three categories—technology-intensive, capital-intensive, and labor-intensive—based on the characteristics of the resources they primarily rely on. This classification allows for a more precise capture of how these differences affect the application of AI Technology. The regression results based on this heterogeneity perspective are presented in

Table 10. Heterogeneity test based on production factors.

	I	II	III
	Model 10	Model 11	Model 12
AI	0.4165	0.6599	0.6948 ***
	(0.3306)	(0.6324)	(0.2642)
Control Variables	YES	YES	YES
Year Fixed Effects	YES	YES	YES
Industry Fixed Effects	YES	YES	YES
N	8942	1732	2192
R2	0.12	0.248	0.214

Standard errors in parentheses. *** p < 0.01.

. The regression analysis reveals that, in technology-intensive enterprises, the impact of AI Technology is not significant, with a coefficient of 0.4165 and a standard error of 0.3306, suggesting that the marginal benefits of AI Technology may be constrained by technological saturation or the challenges of innovation. In capital-intensive enterprises, the coefficient is 0.6599, with a relatively larger standard error of 0.6324. However, it still indicates that AI Technology has a positive effect on innovation performance by enhancing production efficiency, particularly in areas of capital allocation and technological upgrading. For labor-intensive enterprises, the impact of AI Technology on innovation performance is both significant and positive, with a coefficient of 0.6948 and a standard error of 0.2642. This effect is statistically significant at the 1% level, indicating that AI Technology plays a significant role in improving labor productivity, reducing labor costs, and enhancing work efficiency, thereby driving an increase in Corporate Innovation Performance.

Although there are differences in the responses to AI Technology across various types of enterprises, overall, the findings suggest that AI Technology has a universally positive effect on enhancing Corporate Innovation Performance, providing theoretical support for the adoption of new technologies by enterprises in competitive environments.

4.3.2. Heterogeneity Analysis Based on Industry

The application effects of AI Technology may vary across industries. In high-tech industries, technological innovation is a core competitive advantage. Machine learning and data analytics have a positive impact on product development speed and market adaptability in these sectors. Compared to non-high-tech industries, high-tech industries typically have more R&D investment, advanced technological infrastructure, and broader international cooperation networks. As a result, AI Technology can significantly enhance Corporate Innovation Performance by integrating with Organizational Learning Capability in high-tech industries.

To account for the heterogeneity of industry effects, this study refers to the “Statistical Classification of High-tech Industries” and the “Classification Standard for Industrial Enterprises” issued by the National Bureau of Statistics of China. These classifications divide industries into high-tech and non-high-tech sectors. The high-tech category includes industries with high technological intensity, such as information technology, biotechnology, and electronics, while the non-high-tech category covers industries with lower levels of technological investment or innovation output.

The following regression analysis results based on industry heterogeneity are presented in

Table 11. Heterogeneity test based on industry.

	I	II
	Model 13	Model 14
AI	0.2599 ***	0.0261
	(0.0808)	(0.2853)
Control Variables	YES	YES
Year Fixed Effects	YES	YES
Industry Fixed Effects	YES	YES
N	10,663	2290
R2	0.357	0.314

Standard errors in parentheses. *** p < 0.01.

. The results in Table 11 show significant differences in the impact of AI Technology on Corporate Innovation Performance across industries. Model 13, which covers high-tech industries, reveals that the coefficient of AI Technology is 0.2599 and is significant at the 1% level, with a standard error of 0.0808. This indicates that the application of AI Technology can significantly enhance innovation performance in high-tech industries, likely due to the solid technological foundation in these sectors, where firms are more inclined to adopt advanced technologies to maintain their competitive advantage. However, in non-high-tech industries (Model 14), the coefficient of AI Technology is only 0.0261, and it is not statistically significant. This may be attributed to the relatively weak technological foundation in these industries and limited capacity for absorbing and transforming new technologies. In such industries, even though firms may attempt to adopt AI Technology, they may face mismatches between the technology and their existing business processes, resulting in suboptimal application effects.

For corporate managers, understanding the technological application context and innovation potential of their industry is crucial in developing and adjusting technology investment strategies more effectively. This study recognizes that while the classification of industries into high-tech and non-high-tech based on the NBS system provides valuable insights, future research could explore more nuanced classifications that take into account the AI application contexts within traditionally non-high-tech sectors.

4.3.3. Heterogeneity Analysis Based on Firm Size

The scale and development stages of Chinese enterprises vary significantly, with such differences being particularly pronounced in resource allocation, innovation capabilities, and management models. These variations also lead to significant differences in the application of AI Technology, the role of Organizational Learning Capability, and the impact on Corporate Innovation Performance. To address this, this study adopts the heterogeneity perspective of enterprise size, dividing firms into large enterprises and small and medium-sized enterprises (SMEs) for analysis. Specifically, based on the “Standards for Classification of Small and Medium-sized Enterprises”, the sample is grouped, with the results presented in

Table 12. Heterogeneity test based on firm size.

	I	II
	Model 15	Model 16
AI	0.3092	0.5049 *
	(0.2548)	(0.2959)
Control Variables	YES	YES
Year Fixed Effects	YES	YES
Industry Fixed Effects	YES	YES
N	2243	6144
R2	0.373	0.203

Standard errors in parentheses. * p < 0.1.

.

In large enterprises, the coefficient of AI Technology’s impact on innovation performance is 0.3092. Although the significance of this coefficient is relatively low, it still indicates that, despite having abundant resources and a mature management system, the application effect of AI Technology in large enterprises is somewhat limited. In contrast, for SMEs (Model 16), the coefficient of AI Technology is 1.2704 and is significant at the 5% level, indicating that the introduction of AI Technology significantly enhances innovation capabilities in SMEs, even when resources are relatively constrained. This suggests that differences in enterprise scale may stem from disparities in technological infrastructure investment, innovation strategies, and organizational flexibility between large enterprises and SMEs.

In summary, enterprise scale significantly affects the role of AI Technology in improving Corporate Innovation Performance. For policymakers and business decision-makers, understanding the characteristics of enterprise scale and reasonably promoting AI Technology accordingly will help enhance the overall innovation capabilities and market competitiveness of the industry.

4.4. Mechanism Analysis

Table 13. Moderating effect analysis.

	I	II	III
	Model 17	Model 18	Model 19
AI	0.2763 ***	0.2788 ***	0.5469 ***
	(0.0761)	(0.0763)	(0.1804)
Olc		−0.0245 ***	−0.0217 ***
		(0.0066)	(0.0070)
AI*Olc			−0.0011 *
			(0.0006)
Control Variables	YES	YES	YES
Year Fixed Effects	YES	YES	YES
Industry Fixed Effects	YES	YES	YES
N	16,249	16,249	16,249
R2	0.348	0.348	0.348

Standard errors in parentheses. * p < 0.1, *** p < 0.01.

presents the regression results for the moderating effects. First, the results from Model 17 indicate a significant positive correlation between AI Technology and Corporate Innovation Performance (coefficient = 0.2763, significance level = 0.01), suggesting that the application of AI Technology effectively enhances innovation performance, thereby improving the innovation capability and efficiency of firms.

However, Model 18 further shows that when Organizational Learning Capability is considered as a separate explanatory variable, it exhibits a significant negative correlation with innovation performance (coefficient = −0.0245, significance level = 0.01). This result suggests that Organizational Learning Capability does not necessarily directly promote innovation performance; instead, excessive learning capability may negatively impact innovation by increasing decision-making complexity. This finding is consistent with organizational inertia theory, which suggests that firms with high Organizational Learning Capability may become overly focused on internal process optimization, slowing down their ability to adopt and integrate new technologies such as AI.

In Model 19, the interaction term between AI Technology and Organizational Learning Capability is included to examine its moderating effect. The regression results show that the coefficient of the interaction term is −0.0011, significant at the 0.1 level, indicating that Organizational Learning Capability moderates the relationship between AI Technology and innovation performance. When Organizational Learning Capability is higher, the positive impact of AI Technology on innovation performance is weakened. High learning capabilities may lead firms to focus more on process optimization and technological integration, thus slowing the rate at which AI Technology enhances innovation performance.

This finding supports the “Technology-Organization-Environment” (TOE) framework, highlighting the complex relationship between Organizational Learning Capability and technological innovation. According to Teece (2007) [71], a firm’s innovation performance relies not only on the introduction of external technologies but also on the moderating role of internal Organizational Learning Capability. In firms with high Organizational Learning Capability, although AI Technology can improve innovation performance, its effects may be influenced by the existing learning structures and innovation environments, leading to more complex outcomes. Therefore, the moderating role of Organizational Learning Capability in the relationship between AI Technology and innovation performance is of significant importance.

5. Discussion and Conclusions

5.1. Discussion

This study, based on the Dynamic Capabilities Theory, investigates the mechanism through which AI Technology influences Corporate Innovation Performance via Organizational Learning Capability, thereby enriching the existing academic literature. Previous studies have primarily focused on the direct relationship between technological innovation and corporate performance [72]. In contrast, this study reveals that AI Technology indirectly enhances innovation performance by improving Individual Learning Capability and Team Learning Capability, emphasizing the crucial role of dynamic capabilities in technology application. This finding aligns with the Dynamic Capabilities Theory’s assertion regarding the integration and utilization of organizational resources, suggesting that by strengthening Organizational Learning Capability, firms can more effectively leverage AI Technology to enhance innovation performance [71].

Furthermore, this study conducts an in-depth analysis of the multi-level effects of Organizational Learning Capability, highlighting the synergistic impact of Individual Learning Capability and Team Learning Capability in the application of technology. This extends the application of Dynamic Capabilities Theory in the field of technological innovation. Existing research often overlooks the importance of internal organizational learning processes, focusing instead on the technology application itself [73]. This study focuses on how technology influences innovation performance through enhancing Organizational Learning Capability, emphasizing the mediating role of the learning process in technology application. This resonates with Xie et al. (2022) [74] who highlighted the importance of Organizational Learning in technological innovation, further expanding the scope of Dynamic Capabilities Theory in explaining the relationship between technology and performance.

Lastly, the theoretical framework provided by this study offers new perspectives for future empirical research, particularly in the context of the integration of AI Technology and organizational learning mechanisms, thus providing theoretical support for enhancing Corporate Innovation Performance. Unlike prior research [75], this study adopts a dual approach, considering both technology and learning capabilities, while also accounting for background factors such as firm size and industry characteristics, making the findings more comprehensive and practically relevant. Moreover, the study validates the moderating effect of Organizational Learning Capability across different firm characteristics [76], enriching the discussion on the impact of AI Technology on Corporate Innovation Performance and highlighting the critical role of the interaction between Organizational Learning Capability and internal resource allocation in enhancing innovation performance.

5.2. Conclusions

This study, grounded in the Dynamic Capabilities Theory and employing empirical analysis, investigates the impact of AI Technology on the Corporate Innovation Performance of Specialized, Refined, Unique, and Innovative Enterprises in China, along with the underlying mechanisms. The following key conclusions are drawn:

First, there exists a significant positive relationship between AI Technology and Corporate Innovation Performance. Regression analysis shows that AI Technology is positively correlated with innovation performance, and this effect remains significant even after controlling for variables such as firm size, age, and growth potential. This finding is consistent with the research of [35] and further supports the Dynamic Capabilities Theory, which posits that firms can enhance their innovation capabilities and maintain competitive advantage by adopting advanced technologies. Therefore, it is crucial for firms to actively apply AI Technology to drive improvements in innovation performance.

Second, Organizational Learning Capability significantly moderates the impact of AI Technology on Corporate Innovation Performance. Specifically, the stronger the Organizational Learning Capability, the more pronounced the positive effect of AI Technology on innovation performance. In contrast, for firms with weaker Organizational Learning Capability, the impact of AI Technology may be constrained. Therefore, strengthening Organizational Learning Capability, particularly in knowledge absorption, transformation, and application, is essential for firms to fully leverage the potential of AI Technology.

Furthermore, the impact of AI Technology on Corporate Innovation Performance exhibits heterogeneity across factors such as production factors, industry, and firm size. At the level of production factors, enterprises with limited resources or poor technological foundations fail to fully leverage AI Technology, resulting in weaker innovation outcomes [77]. In terms of industry, firms in high-tech sectors are more likely to achieve technological and product innovations, while traditional industries face challenges in applying these technologies. Regarding firm size, large enterprises, with their abundant resources, can more effectively utilize AI Technology and learning resources, leading to greater innovation outcomes. In contrast, small and medium-sized enterprises, constrained by limited resources, fail to fully capitalize on their innovation potential.

Lastly, this study faces limitations in data acquisition and sample completeness. Although the data are derived from Specialized, Refined, Unique, and Innovative Enterprises (SRUIEs) listed in China across various industries, some data related to innovation performance, AI Technology application, and Organizational Learning Capability are missing. Notably, data from non-listed and local enterprises are scarce, limiting the representativeness of the sample. Moreover, the study relies on data from a specific time period. Although it covers the technological innovation phase, the impact of AI Technology and Organizational Learning Capability on innovation performance may vary over time due to technological advancements and changes in the market environment. As a result, data from a single time period may not fully capture this dynamic process. Additionally, the unique policy environment in which SRUIEs operate may limit the generalizability of the findings to non-policy-supported firms or multinational companies. Cultural and institutional differences may also play a significant role in shaping how AI influences innovation performance. Future research should explore these factors by including a broader sample of enterprises from different institutional and cultural contexts, as well as investigating the long-term impact of AI adoption on innovation performance.

Given the limitations of the current study, future research could explore several promising directions: First, a deeper investigation into the differences in the application of AI Technology across industries and its impact mechanisms is warranted. While AI Technology has been widely adopted across various sectors, the specific technical requirements, developmental stages, and competitive environments of these industries vary, leading to potentially significant differences in how AI is applied and its effects. Future research should focus on analyzing these industry-specific differences, exploring how AI adapts in different contexts and how its implementation affects dynamic capabilities such as innovation and market adaptability, thus providing more targeted theoretical support for business decision-making. Second, although Organizational Learning Capability (Olc) is a central component of dynamic capabilities, there remain several under-explored areas within this domain. Future studies should expand the scope to include other dynamic capability variables such as technological innovation capacity, strategic leadership capability, and organizational change capability. These factors may play a crucial role alongside Olc in enabling firms to achieve competitive advantage in a complex environment. In particular, technological innovation capacity is highly relevant to the rapid development of AI Technology and should be explored in future research. Investigating the multi-dimensional characteristics of dynamic capabilities and their interaction in different types of organizations could offer further insights into how these variables collectively enhance a firm’s technological competitiveness and innovation capacity. Lastly, policy support plays an essential role in fostering the application of technology and the development of dynamic capabilities. To enhance firms’ performance in global technological competition, policy frameworks should provide more targeted measures to facilitate the mutual reinforcement of technological innovation and organizational learning capabilities. Future research could explore the mechanisms by which policy support influences these factors across different countries and industries, providing both theoretical and practical guidance for policymakers to support technology-driven enterprises in global markets.