This week Tumblr announced a heavily contested decision, namely banning all sensitive content. A new algorithm was introduced which flags “adult content”. Content flagged as such cannot be uploaded. The algorithm for sensitive content is highly problematic and much discussed by Tumblr users. First, the algorithm does not work properly and often flags things as “sensitive” which are not, leading to hilarious results. This also means that users are complaining and posting pictures of their flagged “sensitive” content at other platforms, such as Twitter.
Second, adult content is an important part of using Tumblr, a platform commonly used by fans to post fan fiction, fan art and more. That is not to say that users upload porn on Tumblr willy-nilly. Overall, it’s a friendly, geeky platform, and hashtags streamline the conversations. Because the content might be unsuitable for certain users and even trigger some, fans tend to label it that way by using hashtags like #nsfw.
In other words, until recently users themselves categorized content, curated it, and handled the gatekeeping. This is all in the hands of a poorly designed algorithm now. The banning of adult content is an extreme decision that truly affects the platform and its culture. Some have even predicted the death of Tumblr. Are they correct? In this post I visualize the different sub-discussions on Twitter around this phenomenon. Because adult content is now banned on Tumblr itself, fans primarily debate the new guidelines on platforms like Twitter.
What I see
For this blog, I scraped recent tweets (1000 posts, retrieved on 05-12-2018) to visualize the debate. Netlytic is a great scraper for visualizing recent Twitter activity. The interface is quite user-friendly once you get the hang of it, and has many options inside the scraper already. You can use this data to visualize name and chain networks which you can import in Gephi as well, a handy open-source visualization tool.
This is a visualization of different communities on Twitter discussing
#tumblr and its new NSFW/adult policies. In the most active, blue circle you see Tumblr’s own Twitter account, which is tweeted at primarily by users criticizing the new guidelines or making jokes about it, displaying their unease. Topics that come up in the other circles: Moving erotic fan content to Twitter or Porn hub, making fun of badly filtered topics/the new algorithm, joking about the death of Tumblr.
As you can perhaps see in the close-up, this graph is sorted on modularity, meaning by density of the network, which allows researchers to visualize different sub-communities and subcultures within selected hashtags or keywords. What you see is that there is not one discussion around Tumblr, but rather different public discourses or spheres.
Some of the sub-groups are just retweeting popular content (e.g. screenshots of funny flagged content that should not have been flagged by the algorithm). The banning of adult content in general is joked about often. Fans call Pornhub a more “fan-friendly” platform than Tumblr, in discussion nodes like this one:
This is the post those users retweeted:
This one also emerged a few times in the network, which jokes about the death of Tumblr but also connects it to cherished fan moments and videos. In a way, it celebrates the good times that (anime) fans had on Tumblr. It shows us quality bishounen and queer content, thereby also poking fun at the new Tumblr rules which censor queer and adult content:
What this means
Tumblr’s new policy and changing culture are characteristic of something bigger. A paradigm shift in which algorithms make decisions for us, platforms get filtered, and corporations and technology take ownership over our creative content. This case is an important one, not the least because Tumblr has been around for a long time and researched in detail. This issue represents important questions that we need to address more in fandom, creative business and scholarship. It goes beyond groups and beyond disciplines. We need to get our heads together in platform studies, data studies, and fan studies to really unpack this. What I’d like to emphasize:
- Platform are not neutral but corporations: While we consider platforms public spaces, they are not. They are corporate entities and in some cases (e.g. Facebook) business empires. The ways in which we connect is in their hands, but should they police their users in this way?
- Algorithms are not perfect and biased: What happens when algorithms, which are currently imperfect and historically grounded, take decisions for us and curate our content?
- The “ideal” online fandom does not exist (yet): What would an open, fan-driven platform look like, and is this the solution? After fanfiction.net changed policies, users migrated to AO3, but I am hesitant whether migration is always the option.
- Platforms capitalize on us: YouTube, Twitch and other platforms capitalize heavily on their core users with ads, by not paying a large part of their content creators, and by automatizing labor. How should we regulate the platform economy as a whole?
- We need access to data: How can users – in a time of datafication and algorithms – regain ownership over their own data and posting behavior in general? Where can we filter, see the data, change procedural rules, and take control again?
We need to discuss these issues, because online platforms are not just technology. They shape our cultures, the way we live them, our entertainment, our politics, our decisions and our world views. To be honest, I am often in doubt whether leaving control of semi-public spaces to the businesses is the best outcome for users. How can we flip this?