The Sentinel Morning Post: Google Morals

Robert Howell

In 2014, the philosopher Robert Howell gave us a surprisingly prescient thought experiment. It goes like this: Imagine you're in New York City and you don't know your way around. And so you open Google Maps to find the best restaurant in town to get a burger. It tells you to follow the blue dot and, hey presto, a delicious burger. But now you open a different app, a new app called Google Morals.

This app allows you to type in any question and it tells you what to do. And so you type in, "Is it better to eat the chicken burger or the vegetarian option instead?" It tells you what to do, and you act upon it. The question that Howell asks is why is it that Google Maps is considered to be okay, but Google Morals is not?

Howell argues that the problem with Google Morals isn't about the answers that it might give, or the reasoning it takes to get there. The problem is what it does to us, the user, because moral reasoning is a practice and a skill. Whenever we wrestle with an ethical dilemma, we are building something (whether you call that character, moral fibre, virtue, conscience, or whatever).

If we outsource that reasoning to an app, we are outsourcing our own moral growth. Howell wrote this in 2014, but today it is here. AI can write better ethical arguments than most people, and we use it to tell us what is right or wrong. But the problem isn't whether ChatGPT is moralising correctly. The problem is whether we should be using ChatGPT to moralise at all.