Barriers to Trust in AI: a Study of the Explainability Technologies Adoption in Banks

With rapidly moving technological progress, especially in the fields of artificial intelligence (AI) and machine learning (ML), new prospects are open for banks and financial institutions. The research aims to understand the barriers to trust in AI by banks. The study is based on qualitative research on drivers of adopting explainability technologies in banks that improve transparency and perception of complex AI algorithms. The paper has identified the main issues related to AI algorithms employment in Eastern European banks and IT AI/ML developers by applying human-centred and ethics-by-design approaches, combined with interviews of the senior stakeholders. The results obtained outline a contrasted view of the AI barriers, and also underline the importance of using AI/ML technologies in banks, the increasing readiness to implement them more extensively, and the challenging aspects in the implementation of AI/ML systems mostly associated with costs and economic efficacy. The study geography was chosen since there is sharp growth in $\mathrm{AI} / \mathrm{ML}$ adoption in Eastern European banks. In each of the surveyed banks, however, “leapfrogging” AI/ML paths had their own drivers and challenges.