Towards Sustainable AI: Green Accounting-Based Model for Measuring the Environmental Cost of Intelligent Systems

The rapid growth of artificial intelligence (AI) has raised questions about its environmental, social, and economic sustainability.This study investigates the influence of three critical factors environmental impact of AI, e-waste and lifecycle implications, and green accounting practices on the development of sustainable AI in Bangladesh.Data were collected through a structured survey and analyzed using structural equation modeling (SEM) with AMOS.The results show that all three constructs significantly predict sustainable AI, with green accounting practices exerting the strongest influence ( = 0.41), followed by environmental impact ( = 0.32) and e-waste considerations ( = 0.27).Together, these predictors explain 65 percent of the variance in sustainable AI, highlighting their collective importance.The findings suggest that addressing energy use and carbon emissions, managing e-waste responsibly, and institutionalizing green accounting are essential steps for promoting sustainable AI.From a theoretical perspective, the study extends sustainability and accounting frameworks into the AI domain.From a practical perspective, it offers guidance for policymakers, industry leaders, and organizations in developing economies.The results contribute to the global discussion on responsible AI by providing empirical evidence from the context of Bangladesh