Infrastructure companies as platform moderations

Hi all,

My friend Suzanne van Geuns and I published an op-ed today for the Brookings Institute, looking at how Internet infrastructure companies engage in content moderation and why their role should be more visible and accountable.

We focus on the role of invisible Internet companies like Cloudflare and how they make decisions about what content stays online. We also talk about their role against the backdrop of covid and the recent BLM protests. We question whether they should have their cake and eat it to by presenting themselves as allies in the fight against racism while wavering to take down alt-right content.

You can find the piece here:

And my tweet about it here:

Would love to hear what people think about the piece and the bigger discussion about privatized moderation of content and hate speech online!




Brilliant and accurate. It resonates strongly with my increasing view that all of these top-down content decisions move the Internet in the wrong direction. And it does not surprise me at all that Cloudfare is not truly forthcoming about its actions, but is rather disingenuous about them. I see a version of this in pretty much all of the big service providers, though I was not aware of CDN’s doing this until I read your essay.

I think that while much of the rationale for such censoring actions get expressed as some sort of service to the public, in fact the purpose of those actions is more in service of those providers maintaining control over their user base, which in turn helps their bottom line. Calling themselves a public utility is feel-good PR because as you say, there can be no true public utility without public oversight.

I do not think the Internet is in a wild west. Those days passed long ago. We are in a period of large-scale top-down control that helps itself by perpetuating an illusion of a wild west.

1 Like

And I think that user control over the content they (we) see is, or should be, a core part of a human-centric Internet. So far it is not. Privacy and user control of their data, yes. User control of the content they/we see, no.

But why can we not be our own content managers? The TCP/IP revolution that made the Internet as we know it possible created a massive network where the intelligence is at the end points instead of the center. So if the intelligence is at the end points, why can that intelligence not control its own user experience? To me, there will never be a human centered internet without this.

Meanwhile this pervasive whack-a-mole process of stifling hate and harmful material goes on, never succeeding.

When the US Supreme Court ruled on obscenity, it allowed local communities to largely define their own standards of what is acceptable. The Internet version of this should mean that we ourselves are the arbiters of our own standards and if I or anyone else objects, we can filter it out. of course that does not go far enough. What is also lacking are well designed and thought-through user tools that help us make such determinations. I predict such tools will never be forthcoming without government requiring it.

Hi John,

Thank you for reading the piece and the kind words, it means a lot. I agree that the Internet is not a Wild West, or at least not as it once was. This turn of phrase is admittedly a little bit of a rhetorical ploy Suzanne and I used. But I do feel that current oversight of infrastructural companies comes down to “cowboy’s honor”.

Whereas the social media companies are subject to more serious oversight, these lower layer companies are subject to less scrutiny. Now, this is by no means to say that government oversight or regulation will improve this situation either. What we saw during the last US Congressional hearings at least, was that much of the debate was fueled by party-politics rather than asserting oversight.

That is not even to speak of a number of countries across the globe who use their power over various Internet companies (through licensing agreements or spectrum access) to make these companies the proxies for their terrible politics.

I like your idea of human control. At the same time, I am worried about who will in effect have the time and other resources to - on top of jobs, families etc - dive into the nitty-gritty of controlling their own user experience at the level of CDNs. We already see that a negligible number of people fiddle with bits they can change. What percentage of people change the details of their DNS resolver? Who checks their CAs? Probably the same (white male) folks currently running companies like Cloudflare or participating in the IETF.

I do not necessarily have a silver bullet. But I do think more can be done by way of regulation, that both forces the cowboys to follow some externally enforced accountability mechanisms without average users needing to become networking experts.

Would love to hear your thoughts!

It is a problem - and it really got going when smart phones came into the picture because masses of people were suddenly on the 'Net without having to do or even know very much. Regardless, here we are. Smart regulation by entities that truly understand what it is they are trying to regulate would get us closer to where we should be (compared to having companies maintain their nanny-state control), but it can never fix what I see as a core problem of the mass spread of disinformation: lack of curiosity as to what is and is not true.

However, while it is political suicide for any elected official to assign blame or responsibility to the electorate, the rest of us can do it. Because what can any society do about people who want to believe things that are demonstrably false? If people have the tools to find out and don’t use them, should the rest of us be penalized? And to me it is a penalty to have been in this business for 34 years and see the tech be so much better but the user experience be so much worse when it comes to individual free choice.

My daughter has kids age 7 and 9. Those kids are quite media savvy for sure, and I believe there is greater burden now on a parent helping their kids navigate through this soup of information and disinformation. But it is still their responsibility more than the state or some corporation. At least I hope they see it that way, because the alternative to my view is a doubling down of today’s problems.

I admit that while I have been thinking about this issue for a long time, my thoughts are still not fully baked. I just know that I am capable of handling my own information inflow, just as I was before these titans of tech assigned for themselves the role of gatekeeper. I don’t need someone shutting off my access to 8Chan or anything else. And I find it inappropriate and offensive that any CDN would curtail that access. Same with Nazi stuff. What if I want to see what they are up to directly? Why should I not be able to do that? Again, the problem to me is not so much that one can see it. The problem is that someone doesn’t know enough, or have enough ways to see that such sites spew hateful bile.

Now designing and creating a comprehensive toolkit for a user to more smartly manage the inflow of content, would be difficult and expensive. But ad hoc censoring is no answer. And while a government can mandate ownership of one’s data and privacy, helping people create and develop a good reliable BS detector is a far taller order. But like I said above, I do not see how any vision of a ‘human centered Internet’ can be realized without it.

I think it would be wrong though to point out these problems if it also gives the impression that I don’t receive benefits from using that same controlled social media. I most certainly do. I have reacquainted myself with a lot of people I grew up with and without the extreme ease of use and low barrier to entry that FaceBook provides, that odds of that happening some other way are close to zero. And I keep up with a lot of people I can’t see at parties and don’t have time to call on the phone and I enjoy the casual party-like feel of most of the discussions I get into there. And I would gladly pay for that service if it gave me the kind of control I want. But I also know that had there been a paywall, most of those same people would not play. Facebook’s great advantage has always been that it largely trades on existing relationships and helps people build and grow that. I wish, along with those benefits, we didn’t get the egomania of the owner, the nanny state vibe, the never knowing what is exactly going on with your data and feeds, and the illusion of control…but I know that there are tradeoffs.

I do lament that there is no meaningful competition for the top tier companies. Maybe there can be no human centered Internet until the fact of anti-trust gets dealt with…