Arising from distributed artificial intelligence (AI) and blockchain (BC), the recently emerging notion of decentralized AI refers to transferring information and learning to various peer-to-peer connected machines which learn based on their locally available data and make decisions individually. Using a decentralized consensus mechanism without the need for trustworthy third parties or intermediaries, decentralized AI provides its users with processes, analytics, and decisions based on trusted, digitally signed, and securely shared data that have been transacted and stored on the BC in a decentralized manner. In order to identify core research focusing on both AI and BC and to find avenues for future research, this study employs thematic analysis of keywords of 1,538 academic publications retrieved from WoS and Scopus databases, as well as bibliometric analysis of authors, affiliations, and sources to examine productivity, citation metrics, and bibliographic coupling. By highlighting the areas of digital transformation, environment/society, decentralized AI, DeFi, and cyber security as the focal points of the BC-AI integration, this paper intends to give researchers a comprehensive view of this convergence and can be used in the industry.