Most people I talk to think that Facebook, Twitter, and other social media companies should take down ugly-but-legal user speech. Platforms are generally applauded for taking down racist posts from the White Nationalist demonstrators in Charlottesville, for example. I see plenty of disagreement about exactly what user-generated content should come down -- breastfeeding images? Passages from Lolita? Passages from Mein Kampf? But few really oppose the basic predicate of these removals: that private companies can and should be arbiters of permissible speech on their platforms.*
At the same time, if the people I talk to are geeky enough, they usually support Net Neutrality. They believe that ISPs, as providers of core Internet infrastructure, should not get to be arbiters of content. ISPs should allow the bits to flow equally -- not suppress or favor particular messages or sources.
This week’s upheaval over the Nazi-inspired Daily Stormer news site raises questions about the spectrum between the two – between ISPs as neutral infrastructure, and platforms as content arbiters. Two of the intermediaries that rejected the Daily Stormer, Go Daddy and Google, were only acting as its domain registrars. They did not host the site, they just made sure that dailystormer.com was associated with the right IP addresses, so users typing the domain and clicking links to it would get there.
So, what do we want domain registrars to be? Speech arbiters or neutral speech facilitators? I’m confident we shouldn’t want the law to force them to arbitrate speech. But I’m honestly not sure about social and moral expectations for them to do so. Here are some considerations.
- Registrars, unlike ISPs, have lots of competition. If one kicks you out, you can find another. That weighs toward letting them choose content.
- Registrars, unlike social media platforms, basically just facilitate the flow of data. They aren’t trying to build communities or algorithmically identify content that users will like. In this sense they feel more like ISPs.
- Most registrars have very few resources to take a good hard look at disputed speech. (GoDaddy and Google are exceptions.) If we ask these small companies to take on content removal obligations, we should not expect nuanced decision-making or robust appeal processes. We should expect legal and important sites from across the political spectrum to go down because someone complained about them.
In conclusion (or non-conclusion), it’s hard. I don’t have a good principle for identifying who, along the spectrum of technical providers online, reasonably “should” make choices about abhorrent-but-legal online speech. I’d be interested to know what engineers who work with end-to-end design principles, and communications law specialists who think about layered models of Internet regulation, have to say. But I definitely think that, even in the midst of our sorrow and outrage over the events in Charlottesville, we should take time think before asking these companies to be arbiters of our speech.
* Maybe more people should oppose it. But that's a topic for another day.