A hybrid two-step approach for assessing the probabilit of training needs on artificial intelligence systems

Artificial Intelligence (AI) represents the core of many technologies and in the last few years, it has become more and more crucial in helping and enhancing decision-making processes. A wide variety of research studies has been developed in AI, covering many different areas, from Health to Agriculture, from Industry to Information Technology. Nevertheless, only a few works have focused on the impact of applying AI on people’s confidence and their reflections on training needs. The novelty of this study concerns the introduction of a hybrid two-step approach based on machine learning and multilevel modeling to assess the effect of people’s awareness, attitude and trust in AI on the probability of training needs. In particular, the Boruta Random Forest algorithm will be applied to identify the key determinants of training needs in AI in eight European countries to be included in the multilevel logit model. Then, the probability of European citizens’ educational needs in AI will be computed and analyzed with respect to gender.