Big data serves as the vital element for machine learning (ML) techniques and artificial intelligence (AI) systems, providing vast quantities of data which is necessary for training and refining algorithms for impellent sharpened decision-making, increasing efficiency and innovation in many businesses including government. Notwithstanding, the rising dependence on information has brought to the very front the basic issues of information protection and security. The fusion of AI and big data has catalyzed breakthroughs on various applications, including natural language processing, recommendation systems, computer vision, and predictive analytics. In this paper, we discussed the privacy and security threats as a main driving force within the ML/AI workflow. Our goal here is to create a bridge between the standardization and research frame to increase the efficiency and consistency of AI applications development which can guarantee user satisfaction while providing a great degree of trustworthiness.