The rapid evolution of artificial intelligence (AI) presents a significant challenge for organizations to keep pace, forcing them to engage in AI sourcing. While extensive research on IT outsourcing for traditional information systems (IS) exists, it remains unclear if established sourcing approaches are suitable to tackle AI sourcing's unique challenges, including ambiguous data ownership, the complex nature of AI models, and their operational uncertainties. This study addresses this research gap by leveraging configuration theory, employing literature reviews and expert interviews with 15 organizations, representing both sourcing providers and customers, to identify unique characteristics of AI-based IS and sourcing particularities. As a result, we identified five key differentiators of AI-based IS with 16 distinctive manifestations and nine particularities associated with AI sourcing. The results enable a systematic understanding of unique AI characteristics and provide avenues for future research to examine the implications of these particularities, e.g., on sourcing strategies.