Modern policing uses artificial intelligence (AI) tools for a variety of purposes, from suspect identification to monitoring to risk assessment. Typically, these tools are thought to be less biased as they use computational methods for policing tasks. How do Americans view these tools, however? And do those evaluations differ depending who supports the use of AI in policing? We field a survey experiment that manipulates AI type and whether the sponsor is a governmental or private entity to explore the contours of support for these tools in the American public. Surprisingly, we find that facial recognition receives more support than general AI tools, particularly among white and Republican respondents. We see little differences in support depending on whether the sponsor is a local police chief or a private entity. This project has implications not only for our understanding of public perceptions of AI in policing, but also for accountability and representation in this essential street-level bureaucracy.