Cornell Computing and Information Science's Jon Kleinberg has been heavily involved in research aimed at applying machine learning to social problems. 

Here, for Harvard Business Review, Kleinberg along with Jens Ludwig and Sendhil Mullainathan present an important and comprehensive "buyer's guide" for anyone who wishes to use data science to create social good. 

There are enormous gains to be made from using machine-learning tools, the researchers say. "But so is the trepidation: as with all new 'products,' there is potential for misuse. How can we maximize the benefits while minimizing the harm?"

What follows is a rich how-to on ways to use machine learning to better society while avoiding common pitfalls.