AI systems are increasingly becoming collaborators rather than merely tools, raising a fundamental question: beyond technical capability, do AI models possess culturally distinct ways of thinking that shape human-AI collaboration? In this talk, I present a programme of research examining AI through the lens of cultural psychology. Drawing on our forthcoming PNAS Nexus paper, I show that large language models developed in different cultural contexts systematically reflect distinct cultural values, norms, and cognitive tendencies, and that cultural alignment between users and AI can enhance creative collaboration. I then present two ongoing studies that extend this work by examining the role of cultural intelligence in human-AI collaboration. The first investigates how users perceive the cultural intelligence of AI models in multicultural contexts. The second examines how human cultural intelligence and perceived AI cultural intelligence jointly shape cross-cultural collaboration. Together, these studies shift the focus from whether AI possesses culturally distinct characteristics to how the cultural capabilities of both humans and AI jointly determine the effectiveness of cross-cultural human-AI collaboration.