Live streaming has emerged as a novel way for product sales, where the streamer plays an important role. However, little is known about how the differences between artificial intelligence (AI) streamers and human streamers impact consumer purchase intentions. Based on social cognitive theory, this study proposed a conceptual model exploring the influences of streamer types, novelty seeking, perceived intimacy, and perceived responsiveness on consumer purchase intention. Based on 387 responses from live streaming users, the results showed that compared with AI streamers, consumers have higher purchase intention when served by human streamers. Consumers perceive human streamers as having more perceived intimacy and perceived responsiveness compared to AI streamers, and perceived intimacy and perceived responsiveness positively affect consumer purchase intention. Furthermore, perceived intimacy and perceived responsiveness mediate the relationship between streamer types and consumer purchase intention. In addition, the moderating effects of novelty seeking on consumer purchase intention are significant. Specifically, lower novelty seeking consumers have higher purchase intentions for human streamers compared to AI streamers, and the mediation effects of perceived responsiveness will be stronger. However, for high novelty seeking consumers, there is no difference in the influence of AI streamers and human streamers on consumer purchase intention, and the mediation role of perceived intimacy will be weaker. This study enriches the theoretical research on AI and live streaming and provides management strategies for live streaming platforms to obtain better benefits through different types of streamers.