Author: Benjamin Mako Hill

Note: I have not published blog posts about my academic papers over the past few years. To ensure that my blog contains a more comprehensive record of my published papers and to surface these for folks who missed them, I will be periodically (re)publishing blog posts about some “older” published projects. This post is closely based on a previously published post by Kaylea Champion on the Community Data Science Blog.

One popular approach to protecting our privacy online is to use the Tor network. Tor protects users from being identified by their IP address, which can be tied to a physical location. However, if you’d like to contribute to Wikipedia using Tor, you’ll run into a problem. Although most IP addresses can edit without an account, Tor users are blocked from editing.

Other research by my team has shown that Wikipedia’s attempt to block Tor is imperfect and that some people have been able to edit despite the ban. As part of this work, we built a dataset of more than 11,000 contributions to Wikipedia via Tor and used quantitative analysis to show that contributions from Tor were of about the same quality as contributions from other new editors and other contributors without accounts. Of course, given the unusual circumstances Tor-based contributors face, we wondered whether a deeper look at the content of their edits might reveal more about their motives and the kinds of contributions they seek to make. Kaylea Champion (then a student, now faculty at UW Bothell) led a qualitative investigation to explore these questions.

Given the challenges of studying anonymity seekers, we designed a novel “forensic” qualitative approach inspired by techniques common in computer security and criminal investigation. We applied this new technique to a sample of 500 editing sessions and categorized each session based on what the editor seemed to be intending to do.

Most of the contributions we found fell into one of the two following categories:

• Many contributions were quotidian attempts to add to the encyclopedia. Tor-based editors added facts, fixed typos, and updated train schedules. There’s no way to know whether these individuals knew they were just getting lucky in their ability to edit or were patiently reloading to evade the ban.
• Second, we found harassing comments and vandalism. Unwelcome conduct is common in online environments, and it is sometimes more common when the likelihood of being identified decreases. Some of the harassing comments we observed were direct responses to being banned as a Tor user.

Although these were most of what we observed, we also found evidence of several types of contributor intent:

• We observed activism: when a contributor tried to draw attention to journalistic accounts of environmental and human rights abuses committed by a mining company, only to have editors linked to the mining company repeatedly remove their edits. Another example involved an editor trying to diminish the influence of proponents of alternative medicine.
• We also observed quality maintenance activities when editors used Wikipedia’s rules on appropriate sourcing to remove personal websites being cited in conspiracy theories.
• We saw edit wars with Tor editors participating in a back-and-forth removal and replacement of content as part of a dispute, in some cases countering the work of an experienced Wikipedia editor whom even other experienced editors had gauged to be biased.
• Finally, we saw Tor-based editors participating in non-article discussions, such as investigations into administrator misconduct and protests against the Wikipedia platform’s mistrust of Tor editors.

In all, these themes led us to reflect on how the risks individuals face when contributing to online communities sometimes diverge from the risks the communities face in accepting their work. Expressing minoritized perspectives, maintaining community standards even when you may be targeted by the rulebreaker, highlighting injustice, or acting as a whistleblower can be very risky for an individual, and may not be possible without privacy protections. Of course, in platforms seeking to support the public good, such knowledge and accountability may be crucial.

This work was published as a paper at CSCW: Kaylea Champion, Nora McDonald, Stephanie Bankes, Joseph Zhang, Rachel Greenstadt, Andrea Forte, and Benjamin Mako Hill. 2019. A Forensic Qualitative Analysis of Contributions to Wikipedia from Anonymity Seeking Users. Proceedings of the ACM on Human-Computer Interaction. 3, CSCW, Article 53 (November 2019), 26 pages. https://doi.org/10.1145/3359155

This project was conducted by Kaylea Champion, Nora McDonald, Stephanie Bankes, Joseph Zhang, Rachel Greenstadt, Andrea Forte, and Benjamin Mako Hill. This work was supported by the National Science Foundation (awards CNS-1703736 and CNS-1703049) and included the work of two undergraduates supported through an NSF REU supplement.

Note: I have not published blog posts about my academic papers over the past few years. To ensure that my blog contains a more comprehensive record of my published papers and to surface these for folks who missed them, I will be periodically (re)publishing blog posts about some “older” published projects. This particular post is closely based on a previously published post by Nate TeBlunthuis from the Community Data Science Blog.

Many online platforms are adopting AI and machine learning as a tool to maintain order and high-quality information in the face of massive influxes of user-generated content. Of course, AI algorithms can be inaccurate, biased, or unfair. How do signals from AI predictions shape the fairness of online content moderation? How can we measure an algorithmic flagging system’s effects?

In our paper published at CSCW, Nate TeBlunthuis, together with myself and Aaron Halfaker, analyzed the RCFilters system: an add-on to Wikipedia that highlights and filters edits that a machine learning algorithm called ORES identifies as likely to be damaging to Wikipedia. This system has been deployed on large Wikipedia language editions and is similar to other algorithmic flagging systems that are becoming increasingly widespread. Our work measures the causal effect of being flagged in the RCFilters user interface.

Our work takes advantage of the fact that RCFilters, like many algorithmic flagging systems, create discontinuities in the relationship between the probability that a moderator should take action and whether a moderator actually does. This happens because the output of machine learning systems like ORES is typically a continuous score (in RCFilters, an estimated probability that a Wikipedia edit is damaging), while the flags (in RCFilters, the yellow, orange, or red highlights) are either on or off and are triggered when the score crosses some arbitrary threshold. As a result, edits slightly above the threshold are both more visible to moderators and appear more likely to be damaging than edits slightly below. Even though edits on either side of the threshold have virtually the same likelihood of truly being damaging, the flagged edits are substantially more likely to be reverted. This fact lets us use a method called regression discontinuity to make causal estimates of the effect of being flagged in RCFilters.

To understand how this system may affect the fairness of Wikipedia moderation, we estimate the effects of flagging on edits on different groups of editors. Comparing the magnitude of these estimates lets us measure how flagging is associated with several different definitions of fairness. Surprisingly, we found evidence that these flags improved fairness for categories of editors that have been widely perceived as troublesome—particularly unregistered (anonymous) editors. This occurred because flagging has a much stronger effect on edits by the registered than on edits by the unregistered.

We believe that our results are driven by the fact that algorithmic flags are especially helpful for finding damage that can’t be easily detected otherwise. Wikipedia moderators can see the editor’s registration status in the recent changes, watchlists, and edit history. Because unregistered editors are often troublesome, Wikipedia moderators’ attention is often focused on their contributions, with or without algorithmic flags. Algorithmic flags make damage by registered editors (in addition to unregistered editors) much more detectable to moderators and so help moderators focus on damage overall, not just damage by suspicious editors. As a result, the algorithmic flagging system decreases the bias that moderators have against unregistered editors.

This finding is particularly surprising because the ORES algorithm we analyzed was itself demonstrably biased against unregistered editors (i.e., the algorithm tended to greatly overestimate the probability that edits by these editors were damaging). Despite the fact that the algorithms were biased, their introduction could still lead to less biased outcomes overall.

Our work shows that although it is important to design predictive algorithms to avoid such biases, it is equally important to study fairness at the level of the broader sociotechnical system. Since we first published a preprint of our paper, a follow-up piece by Leijie Wang and Haiyi Zhu replicated much of our work and showed that differences between different Wikipedia communities may be another important factor driving the effect of the system. Overall, this work suggests that social signals and social context can interact with algorithmic signals, and together these can influence behavior in important and unexpected ways.

The full citation for the paper is: TeBlunthuis, Nathan, Benjamin Mako Hill, and Aaron Halfaker. 2021. “Effects of Algorithmic Flagging on Fairness: Quasi-Experimental Evidence from Wikipedia.” Proceedings of the ACM on Human-Computer Interaction 5 (CSCW): 56:1-56:27. https://doi.org/10.1145/3449130.

We have also released replication materials for the paper, including all the data and code used to conduct the analysis and compile the paper itself.