Getting Past Logic

To the legal formalist, the law consists in applying rules to facts using logic, similar to working out a mathematical proof. It is a common view of computers that they should be thought of in the same way: that they run algorithms, which should be described using rules of logic. This view of computing is appealing to formalist lawyers, who see the link between statutory law and algorithmic source code. But just as Holmes argued against the formalist view of law, we must let go of algorithmic thinking if we are to make sense of machine learning. The essential parts of machine learning—the training dataset, and the learnt parameters—cannot be understood by looking at source code. True, it does use algorithms to learn those parameter values, most notably the Gradient Descent algorithm. But to focus on algorithms and logic is to miss what is distinctive about machine learning. It leads to misguided views about regulation and control.