{"id":73986,"date":"2021-09-28T15:43:07","date_gmt":"2021-09-28T15:43:07","guid":{"rendered":"https:\/\/80000hours.org\/?post_type=article&p=73986"},"modified":"2024-03-21T21:10:45","modified_gmt":"2024-03-21T21:10:45","slug":"moral-uncertainty","status":"publish","type":"article","link":"https:\/\/80000hours.org\/articles\/moral-uncertainty\/","title":{"rendered":"Moral uncertainty: how to act when you\u2019re uncertain about what\u2019s good"},"content":{"rendered":"

We can be uncertain about matters of fact, like whether it’ll rain tomorrow, but we can also be uncertain about moral claims, like how much to value the interests of future generations<\/a>.<\/p>\n

In the last decade, there has been more study of how to act when uncertain about what’s of value, and our cofounder, Will MacAskill, has written a book<\/a> on the topic.<\/p>\n

An approach that’s common is to pick the view that seems most plausible to you, and to go with that. This has been called the ‘my favourite theory’ approach.<\/p>\n

But this approach seems bad. Consider a situation like this:<\/p>\n

You’re at a restaurant and can order either foie gras or vegetarian risotto. You think there’s a 55% chance that animal welfare has no moral significance, and a 45% chance that it does, which would mean it’s deeply wrong to eat the foie gras. Personally, you’d find either meal equally delicious.<\/p>\n

The favourite theory approach would say you should act as if animal welfare doesn’t matter, and so it’s equally good to order either the foie gras or the risotto. But it seems clearly better to pick the risotto.<\/p>\n

Rather than go with your favourite theory, we think a more plausible approach is to consider a range of perspectives, and take the actions that seem best on balance.<\/p>\n

Exactly how to determine which actions seem best ‘on balance’ is highly debated, but we think one likely consequence is a type of moral caution — if one plausible perspective says an action is extremely wrong, we should probably not take the action, even if other perspectives say it’s permissible or even good.<\/p>\n

Some effects this has on our advice:<\/p>\n