Ethical Technology and Political Strategy
The Case Against Political Blocks in Software and Networks
One of the benefits of having a blog is that if I keep finding myself in the same debate, I don't have to waste my time explaining the same points again and again. Instead, I can lay out my arguments in detail in a blog piece, then link to it with a TL;DR. That way, people who want to debate my actual position (rather than a strawman) can click the link and have a read, before continuing the discussion.
In 2020, CoActivate.org - my generous blog host for over a decade - went down never to return. Fortunately, thanks to the WayBack Machine at Archive.org, many of the Disintermedia blog pieces I published there still exist. But they can be much harder to find. So as with the piece on the downsides of HTML email, I'm going to republish them here as the subjects they cover come up again.
This one originally appeared in November 2019 under the title The Case Against Political Blocks in Software and Networks. I republished it on my personal blog three years later, under the title Political Strategy and Ethical Technology. The event it was responding to turned out to be a bit of a storm in a teacup. But the larger issue of how left/right politics interacts with the development of ethical technology remains a tricky one, and I still stand by what I said about it, which was…
There’s been a lot of wailing and gnashing of teeth over the last few months, about how to deal with an infamous, far right social network site. Before I say anything else, I want to make it clear that my politics have been staunchly leftist and anti-racist since my childhood. I find the politics of the site in question appalling, and I have no intention of promoting their toxic brand, even when criticizing them, so I will refer to it here as Cess.pit.
A little background. As my three regular readers will know, I’m an excitable champion of the federated social web, the fediverse to its fedizens. Unlike users in centralized platforms like FarceBook or Titter, users of the fediverse join an “instance” - a server running web software compatible with fediverse protocols like ActivityPub - but they can follow and communicate with users on other instances, run by completely different people. A few months ago, Cess.pit moved its users to a fork of Mastodon, one of the most popular of the various software projects that use ActivityPub, potentially allowing them to recruit from and harass users on other instances.
The admins of many instances responded to news of this by adding cess.pit to their blocklists. This isn’t unusual. The servers that host our email use blocklists all the time against mailservers that send lots of spam or virus attachments. I believe that the power to block users and instances ultimately belongs in the hands of the users of social network software, but given the current state of federated network software, I’m fine with instance blocks. So long as they’re being used by admins to prevent things like spamming, flooding, and harassment, for the benefit of their users, not to police who their users can communicate with, for ideological reasons those users may not agree with.
I think it’s worth pointing out that treating all Cess.pit users as if they were all committed fascist organizers does have downsides. A lot of young or politically naive people wander into online spaces like Cess.pit without really understanding how deep those particular rabbit holes go, and these are the “prospects” the actual fascists hope to indoctrinate and recruit. If Cess.pit is federated with at least some instances that aren’t full of fascists and sympathizers, prospects will be much more likely to get access to other perspectives, making it more likely that they will broaden their minds, realize who they’re hanging out with, and get out. If the only people users on Cess.pit can talk to is each other, chances are that actual fascists will find indoctrination and recruitment much easier [note: in the end Cess.pit stopped federating with other instances, perhaps they too came to this obvious conclusion?].
The other thing that happened is that the developers of Tusky, one of the mobile apps that can be used to connect to instances of Mastodon (and other fediverse software that uses the same system for communication between apps and servers), decided to add a blocklist to their code and add Cess.pit to it. I wasn’t the only person who was sympathetic to their reasons for doing this, but concerned about the possible unintended consequences. Let’s unpack that a bit.
Let’s say a Bad Actor wants to shut down queer leftist GroupX and stop them communicating online. They publicly smear them as “terrorists” or “spreading kiddy porn” or whatever, and start approaching hosting platforms and software developers, demanding they take action to stop GroupX using their tools.
Let’s say they approach Mozilla about building anti-GroupX blocks into Firefox, during the time when Brendon Eich was CEO of Mozilla’s commercial arm. Brendan is going to say “no”. Because even though he’s right-leaning and may well disagree strongly with the politics of GroupX (which is why he was forced out of the CEO role), there is a longstanding principle in internet tech that we don’t implement political blocks at the level of code and network protocols. Those are decisions to be made autonomously by users (network or “bottom up” decision-making), not centrally by engineers or system administrators (pyramid or “top-down” decision-making).
The principle does mean, in theory, that free code developed by leftist anarchists could end up getting used in some way by fascists. But unless you empower the state to maintain a register of fascists and stop them using the net at all, it’s unavoidable that they are indirectly using all sorts of free code, developed by people from all sorts of backgrounds, who have all sorts of reasons to be horrified by the idea of fascists using their code.
When you stop and think about it though, its obvious that the opposite is going to happen much more often. Radical leftists do not develop the majority of free code software. Every day we are directly and indirectly using all sorts of free code, developed by people from all sorts of backgrounds, who have all sorts of reasons to be horrified by the idea of us using their code. But they don’t try to stop us, for the same reason I believe Brendan Eich would have said “no” in the hypothetical example above. They understand that the whole purpose of defending software freedoms is to prevent powerful groups from using their monopoly on the money and infrastructure that funds most software development, to enforce their politics on its users.
So what happens when developers start implementing political blocks at the code level, as Tusky did? Do the ends justify the means?
In the short term, even if every fediverse app followed the Tusky example, life becomes mildly inconvenient for the Cess.pit folks, who can just start distributing their own forks of their preferred apps from their own websites. In fact, they’ve now set up their own app store. No real harm been done to their operations. But in the long term, it starts to normalize the idea that its OK to use the roles of developer, engineer, or hosting provider, to police other people’s politics.
Imagine if all the technical folks who disagree with radical left views started doing the same things to us, that some of us have done to the users of Cess.pit. Imagine centrist liberal CEOs at Mozilla, Goggle, Apple, and Microsoft, building blocklists of radical left websites and media outlets into Firefox, Chrome, Safari, and Edge. Imagine we had to develop all our own software and host all our own infrastructure, and try to convince other users to only use fringe apps and hosting platforms that don’t have built in censorship of our politics. This is incredibly threatening to the radical left - far more threatening than Cess.pit itself.
This is why a lot of people in the ethical tech community disagreed with the Tusky dev’s decisions, even though we respected their right to make it, while the developers of other apps decided instead to respect the longstanding principle of political neutrality discussed above. This is why many of us were horrified when developers who chose not to hardcode blocks into their apps became the target of coercion by some radical leftists. Like those who tried to get Fedilab removed from app stores by claiming its developers “explicitly chose to enable hatred and violence through their app”.
These smear campaigns and false accusations are appalling in themselves, exactly the tactics of the Bad Actors I described above. But even worse, to the degree that they are successful in the short term, they risk massive damage to the prospects of the radical left in the longer term, for reasons described above. They also do massive damage to the sense of goodwill and common cause that convinces people who could earn massive salaries in the corporate tech industry, to spend their spare time - or take more precarious jobs - to write free code that benefits all human beings. Including radical leftists.
… and yes, sometimes including fascists. But for the reasons discussed above, I think the benefits of that far outweigh the costs.
Photo by Jean-Pierre Dalbéra, used with thanks under CC BY 2.0