The Computer Judge
What mutton-headed, technologically myopic luddite said this? I confess that these are my own words, as they appeared in 1986 in the Modern Law Review. Although this was comfortably more than thirty years ago, I can recall quite vividly what was going through my head (for want of a better term) when I wrote that passage. Today, I disagree with much that I said then. Emotionally, I no longer have any sense of horror in contemplating the possibility that judges might roundly be outperformed by machines. Technically, the passage of time has put me out of date. Computers often can (in some constrained circumstances) satisfactorily process speech and natural language. I also failed (along with most computer scientists) to predict that many of the remarkable advances in computing would come not through explicitly programming systems (whether, for example, to exhibit political preferences or creativity) but through machines ‘learning’ from vast sets of accumulated data. Morally, when I spoke of the values of western liberal democracy, I was reflecting the mood of the late twentieth century. As technology advances, it transpires, as Jamie Susskind explains in Future Politics, that our political conceptions change too. Liberal democracy in the twenty-first century may be significantly different from its ancestor.