Wednesday, September 25, 2013

What is "Effective Altruism"?

Effective altruism is altruism that attempts to be maximally effective. An effective altruist chooses her charitable actions not, for instance, by thinking of what she's most passionate about (helping the homeless, say) and then taking the most obvious, readily available actions, or those with the most emotional impact for her (like volunteering at a soup kitchen), but instead tries to take into account all the possible ways she might have of affecting the world, and then picking the one she expects, based on empirical evidence and careful thought, to cause the most good.

Some people's answer to, "What's the most good I can do?" is "Donate to the charity that's most cost-effective at saving lives." (It's orders of magnitude more effective to donate to the Against Malaria Foundation than to donate to your local children's hospital, for example.) Some take non-human animals into account and end up with very different kinds of answers, as you can imagine given how many more non-humans there are than humans. Some think the answer is to choose a career that will make them as much money as possible while doing relatively little harm so they'll eventually have lots more money with which to do good. Some think preventing the sudden extinction of all of humanity ("reducing existential risk") is more effectively altruistic than any specific short-term act of charity we could perform for an individual.

The effective altruists focused on the far distant future are those who value future people as much as current people, and who believe that the distant future is likely to to contain vastly more people than the present. This is the category I fall into. A focus on the far distant future can look even less like conventional charity, because some methods of affecting the future of humanity can be extremely unintuitive. Some people intend to eradicate death caused by aging. One organization called the Machine Intelligence Research Institute intends to build and artificial intelligence that understands and cares primarily about human values, can modify its own code in order to get better at effective altruism, and by this means ends up way better than any human could ever be at making the entire future awesome.

So effective altruism is extremely diverse. It can be giving a dollar directly to someone who needs it, or it can be researching how to teach a computer what humans care about. What matters is the motivation: effective altruists want to be effective, and they try to know something about how to do that.

A few organizations in what's recently become know as "the EA movement":

For a somewhat more in-depth look at the EA movement, check out lukeprog's summary on, written shortly after the very first Effective Altruism Summit in July, 2013. Or see Peter Singer's Ted Talk on the topic if you're tired of reading.