Despite the increasing collaboration with human and AI to accomplish complex goals as a team, there still remains a research gap about which communication style facilitates team performance and team satisfaction in Human-AI Teaming (HAIT). However, communication styles of AI in human-AI teaming and their impacts on team effectiveness remain understudied. To address this issue, we aim to explore how AI communication style (AICS) and the user’s cognitive elaboration (CE) during the user of AI improve the outcomes of HAIT. Results from a 2 (AICS: social-oriented vs. task-oriented) X 2 (CE: mindful use vs. mindless use) between-subjects experiment with 89 participants show that interacting with social-oriented communication style of AI enhanced anthropomorphism, which led to team viability and team identification. In addition, combination of AICS and CE increased direct request and ultimately increased both team viability and team identification. Our study contributes to chatbot designs in teaming context by providing theoretical and practical implications to design appropriate communication style of AI.