With the development of artificial intelligence (AI), a variety of opportunities have emerged to implement machine learning solutions to real world problems. However, there is also an even increasing need to make sure that these techniques perform as intended so different industries where trust plays an important role could also benefit from machine learning solutions. Responsible AI, as a new paradigm tries to fill this gap. However, still less attention has been put to investigate the completeness of its layers. Recently, some studies started to focus on the environmental impact of machine learning algorithms that shall also be part of this AI framework. In this research we try to investigate the potential gaps in the current responsible AI paradigm and propose a sustainable AI framework on a machine learning-based credit scoring case study. We focus on the bias, explainability and environmental impact of these algorithms in one analysis that was mainly investigated separately in other research papers. It will help groups in society to get access to credit that would have been otherwise ignored by traditional credit scoring firms, increase transparency as it will not consider AI as a black box and measure environmental impact that could also highlight inefficiencies while achieving climate change goals.