Digital transformation and environmentally sustainable innovation: Based on machine learning and text analysis...

As global environmental challenges intensify, environmentally sustainable innovation (ESI) has emerged as a critical factor in firms' pursuit of green development strategies. Although traditional literature has emphasized the potential value of digital technologies, less attention has been paid to the transformative role of both internal and external digital transformation in driving ESI. We employ the bidirectional encoder representations from transformers (BERT) and bag of words (BoW) models to construct the ESI and digital transformation variables, facilitating an in-depth examination of the impact of digital transformation on environmental sustainability. Based on Chinese listed company samples, the results show that: first, there is a negative relationship between internal digital transformation (IDT) and ESI; second, external digital transformation (EDT) positively impacts ESI. Further analysis shows that institutional environment and resource endowment simultaneously weaken (strengthen) the negative (positive) relationship between internal (external) digital transformation and ESI. We shed light on the mechanisms through which digital strategies advance environmental innovation, offering important foundations to inform policy and guide sustainable development practice.