AI, Ethics, and the Law

Abstract

There are several ways in which AI can cause ethical problems in working life. Taking the field of law as an example, there are three key ways in which using algorithmic solutions without careful oversight can cause problems. First, systems can be gamed. Second, by using human data to train on, systems can reproduce human prejudices and bias. Third, prediction algorithms can become self-fulfilling as their effect on reality feeds back to them, producing a loop. Most importantly, though, use of AI tends to stagnate our ethics.