»Tools are for the worker«

Remarkable recent advances in machine learning or artificial intelligence call for careful consideration of the nature of this technology and how we use and relate to it. I propose that the language of 'instrument' and 'instrumental cause,' drawn from scholastic philosophy, offers promise, especially as set out by Thomas Aquinas and developed in later Thomism. This tradition is not dismissive of instruments, recognising that there are things we can do with them that we cannot without, but it retains an emphasis on human beings as the makers and users of these instruments. It belongs to the definition of an instrument that it acts only because it is moved to action by an agent. I set out six features of instruments, drawing on this tradition, and use them to refine an understanding of machine learning, with human makers and users in primary position. I discuss some additional distinctions between forms of instrumental causation that are fruitfully suggested by scholastic thought. I conclude with Aquinas's contention that »tools are for the worker,« and should be treated as means, deployed to serve human ends.