Patterns and antipatterns, principles, and pitfalls: Accountability and transparency in artificial intelligence

Abstract

As researchers, we love to envision the real-world scenarios in which our research could offer recipes for improving the world. However, the truth is that powerful technologies are rarely used for good only, and AI is no exception. Although envisioning the ways in which our work could unintentionally lead to harm is not as enjoyable, this article has presented several ideas and suggestions for promoting a culture of AI research in which researchers can play as active a role in controlling the potential misuse of AI as they do in advancing its potential for good.