Nadia Eghbal

<-- home

Moral infrastructure

I was a vegetarian for ten years. When I was twelve, I read something on the internet about how poorly animals were treated on factory farms, and I decided that it was unethical for me to continue eating meat.

Because I was vegetarian for ethical reasons, I was determined not to do anything that would make me a hypocrite. I didn’t eat fish (oddly, there are a lot of vegetarians who do!). I didn’t wear leather or fur. I didn’t eat gelatin. Anything that contributed to the death of animals, I didn’t want to participate in.

One day, I came across a comic about how farming vegetables kills field mice. Heavy machinery rolling through giant open fields, not to mention toxic fertilizer and pesticides: you do the math. “Hmm,” I thought. “I guess that doesn’t make me a very good vegetarian.” But I wasn’t sure what the alternative was. Not eating vegetables?

Over the years, I wrestled with more vegetarian koans like these. Tires are made from animal products. Beer and wine are made with fish bladder. Medication can contain gelatin.

And then there were the inevitable riddles from curious friends, things like: “If the animal is already dead, would you eat it?” “If you order food that accidentally has meat in it, isn’t it more wasteful to just throw it away?” My answers to these questions were “no”, but I realized that I was saying no not because they broke my moral framework, but because I wanted to uphold the idea of ethics-based vegetarianism. It was about maintaining my identity as a vegetarian, rather than evaluating the question on its own merit.

Eventually, I started eating meat again. I concluded that there was no way to be a perfect vegetarian, so I had a few options: 1) make myself crazy by trying to cut every tie to every animal death I could think of, 2) accept the inherently imperfect logic of vegetarianism and strive for “good enough”, or 3) just let it go. Diets are personal, so while I respect anyone’s decision to be a vegetarian, I decided I wasn’t comfortable ascribing to a theory that was impossible to fully practice.

Lately, I’ve noticed that every time something awful happens on the internet, we start blaming infrastructure providers. I understand the intention behind this. Like my twelve year-old self reading about the mistreatment of animals, we’re horrified by the bad things that happen in this world. These stories make us feel a profound loss of control, and we process them by finding a way to believe that we could prevent anything like it from happening again.

Translating this theory to practice, however, is a Sisyphean endeavor. It starts with cloud hosting and payment providers. What about the programming languages, frameworks, and libraries that were used to build these websites? Should Gmail ban all hateful people from using its service to communicate? Should Apple not sell them iPhones? What about the food that nourishes and sustains these people, or the clothing brands they wore in the act? How far are we willing to go to de-platform bad actors?

There’s a reason why open source licenses specify that the code “must not discriminate against any person or group of persons”, nor “restrict anyone from making use of the program in a specific field of endeavor”. It’s not because they actually condone everything people might do with the code. It’s because they recognized it’s impossible to attach instructions to the invention and spread of technology, and they had the foresight to protect developers from liability. If that’s one end of the spectrum, what’s the other end, and where exactly do service providers fall?

Technology itself may be amoral, but what we do with technology is absolutely up for moral debate. But we should approach that conversation carefully. Otherwise, like the vegetarian koans, what starts out reasonably plausible will quickly devolve into a series of what-abouts. [1]


[1] Nabeel asked a question on Twitter that helped me clarify my position, which was, is my premise here that “because it is impossible to be 100% good, it isn’t worth being 75% good”? My response was that it’s more like “it’s impossible to be 100% good”, so, like vegetarianism, where we draw the line is not immediately obvious. Although I tend to prefer striving for a separation between infrastructure and how it is used, I think there are some very good arguments for why providers should get involved…but the point I hoped to make is that wherever we land is a matter of policy, not ethics.