An ever growing problem of the modern social media-rich world is misinformation. The trust that was previously placed into government officials and journalism has eroded; the internet gave everyone a voice but with it made it so much more difficult to distinguish truth from fabrication. The consequences of this are very real: Almost eradicated illnesses are making a comeback because people refuse to vaccinate their children, Covid-19 is continuing to spread because people refuse to wear masks and practice social distancing, more and more people start believing that the Earth is flat and descend down a rabbit hole of ever more absurd conspiracy theories.
The social media giants have acknowledged the problem: Both Facebook and Twitter are taking measures to try and limit the damage of misinformation. Both take the fact check approach, wherein a dubious claim that has attracted enough attention on the platform is disputed by deferring to one or multiple trusted authorities. It is a step in the right direction but we must consider how well it would fit into the decentralized model, which is what we’re working with. In both cases, Twitter and Facebook unilaterally decide a) which claims deserve a fact check and b) which fact checking authorities to defer to. Facebook has already gotten in trouble for picking some very dubious fact checkers.
So we have issues on two layers: The fact checkers selected by the platform may not be the ones that the users actually trust, and only claims that the platform decides to fact check get any treatment. On a decentralized social media platform like Mastodon, there is no central authority that can make those decisions, and while you may argue that its more localized governance structure (where a server’s admins and moderators have fewer users to take care of and users have the freedom to pick the server that fits their needs the best) would be an improvement over this, there is a practical limit to how much micromanagement we can expect independent admins and moderators to perform.
While we routinely observe blatant conspiracy theorists being kicked off well-moderated Mastodon servers, the often volunteer staff simply cannot monitor every message for misinformation and link it up with appropriate resources. For the same reason we oppose various upload filter initiatives – manually checking every message on social media does not scale and any automation is so complicated that it inevitably leads to centralized solutions that are equally inaccessible for small players. Regardless, the takeaway is, if we want to tackle misinformation on decentralized social media, we need a solution that does not rely on manual action by server staff.
In late 2018 I was approached by someone from University of Greenwich who wanted to investigate potential solutions to this problem and wanted my advice, support, and knowledge of decentralized social networks. It was an invitation to participate in an academic research project EUNOMIA with, among others, three different universities (University of Greenwich, University of Nicosia, and University of West Attica) and a grant from the European Commission’s Horizon 2020 program – an extremely flattering acknowledgement of Mastodon’s importance. Indeed, Mastodon was the perfect choice for this purpose: An extremely easy to use, well-documented, and extensive API that not only allows, but encourages the creation of alternative user interfaces; and the ability to essentially run a fully-featured social network in an entirely sandboxed environment.
What EUNOMIA aims to be at the end of its 3-year development road map is a “digital companion” – in essence, an alternative user interface, containing a toolkit that would facilitate the discerning of manipulated or incorrect information. Facilitation is key, here: The user would be the ultimate authority for making a call on what they trust or distrust, what EUNOMIA would provide is easier access to the kind of criteria the user deems important for that decision. Someone might want to be notified if a post uses manipulative wording to distort a claim, someone else might want to see if similar messages have been posted by other people before and the one that you see is less accurate, other people may want to check with the wisdom of the crowd and pay extra heed when lots of people distrust a message. Any one method is imperfect by itself, but in tandem they may make fact checking more accessible.
The EUNOMIA “digital companion” is built on Mastodon but they are two completely separate projects. If you would like to follow EUNOMIA’s progress and provide any feedback, please follow its Mastodon account: @Eunomia@mastodon.social