The rapid progress of artificial intelligence (AI) raises hopes and fears about its future and the effect it will have on human societies. These hopes and fears depend upon people's conception of human nature, anticipated impact of technology on society, and expectations of how AI might align with the best or worst aspects of human moral behavior. Considering the immediate, near term, and more future-oriented social implications of AI raises many ethical issues about how to use and regulate AI technology. These considerations also raise new possibilities in how AI might engage with people. I describe one approach to improve human flourishing by explicitly making AI more compassionate with application to healthcare. Rather than only trying to prevent harms from technology, a constructive approach to compassionate AI can benefit society and define a target for AI development that also addresses concerns about how AI might behave in the future.