Artificial intelligence (AI), as a new technology, not only revolutionizes economic development, but also provides an opportunity for environment governance. Extant studies primarily explore the environmental performance of AI from a macro perspective, while evidence on how AI technology applications affect firms’ energy-saving behavior is scarce. Employing Python technology to recognize AI-related keywords in the annual reports of listed enterprises and adopting data on corporate energy consumption from 2011 to 2020, we explore the impact of AI on corporate energy consumption intensity (CECI) and its mechanisms. We observe that AI technology applications reduce CECI. After a range of robustness tests, the conclusions are still solid. The mechanism analysis reveals that AI cuts CECI through spurring firm green innovation, stimulating firms to introduce new equipment, and reducing firms’ internal management costs. Heterogeneity analysis reveals that this negative impact is more prominent for SOEs and private enterprises’ energy intensity; we also find that this effect is more pronounced for high-tech industry enterprises and high-polluting enterprises. Our findings provide micro evidence for policymakers to reduce corporate energy intensity and realize energy conservation and emission abatement targets.

中文翻译：

---

人工智能技术应用对企业能源消费强度的影响


人工智能 （AI） 作为一项新技术，不仅彻底改变了经济发展，也为环境治理提供了机会。现有的研究主要从宏观角度探讨人工智能的环境绩效，而关于人工智能技术应用如何影响公司节能行为的证据很少。利用 Python 技术识别上市企业年报中与 AI 相关的关键词，并采用 2011 年至 2020 年的企业能耗数据，探讨了 AI 对企业能耗强度 （CECI） 的影响及其机制。我们观察到 AI 技术应用降低了 CECI。经过一系列稳健性测试，结论仍然是可靠的。机制分析表明，人工智能通过刺激企业绿色创新、刺激企业引进新设备以及降低企业内部管理成本来削减 CECI。异质性分析表明，这种负面影响对国有企业和民营企业的能源强度更为突出;我们还发现，这种影响对于高科技产业企业和高污染企业更为明显。我们的研究结果为政策制定者降低了企业能源强度并实现了节能减排目标提供了微观证据。