Abstract
In the context of increasing environmental and social responsibility concerns, organizations are looking for innovative methods to assess and improve sustainability performance. This study explores the role of artificial intelligence (AI) – neural networks, in developing a probability-based evaluation system for organizational sustainability score. Traditional evaluation methods frequently rely on pre-established performance indicators, which can introduce subjectivity and inaccuracies. To overcome these limitations, the research proposes a neural network model that integrates economic, social, and environmental dimensions into a structured evaluation framework. The study uses data collected from 30 companies listed on the Bucharest Stock Exchange. The neural network was developed in MATLAB, with a feed-forward structure with two hidden layers and a Levenberg-Marquardt training algorithm. The results – probabilistic organizational sustainability score reflects an intermediate position associated to the analyzed companies, indicating therefore an acceptable compliance level, but also the existence of improvement opportunities; likewise environmental and social dimensions have a stronger influence on organizational sustainability, while the economic dimension, although relevant, has a lower impact. The findings demonstrate that AI-based models offer a more dynamic and objective approach to sustainability assessment, reducing human error and improving the accuracy of predictions. The research contributes to the literature by introducing a structured, data-driven methodology, providing valuable insights for organizational managers and researchers interested in AI-assisted decision-making.