I started following Wikipedia community organization from 2012. At that time a trend was beginning to talk about addressing online harassment, especially harassment of women by men, but at the time each instance of the problem was imagined as an independent anomaly rather than a systemic problem. By that I mean that big online communities like Facebook, Twitter, Reddit, Wikipedia, and all the others did not organize public discussions about the systemic causes of online harassment as a consequence of the nature of online social interaction. Now, five years later, things have changed and it is routine for everyone managing online communities to have some shared background context, mutual understanding, and objective distance for discussing the problem. One of the stumbling blocks that had to be overcome recently was the recognition that online sociology with harassment plays out differently than offline sociology. This is true for many practical matters of online social interaction, but the digital age as not yet matured, and we are still having to process many broad categories of social interaction and talk through how they differ when done online versus offline. Another of the barriers was mutual acknowledgement that all online communities have this problem. Before 2012 and perhaps not before 2014-15 it was not common knowledge that all online communities had this problem, so the public-facing corporate representatives of every individual online community took action to suppress the fact that their community experienced harassment problems. The logic was that if any online community stepped up and announced that they had a major harassment problem that was unlike anything in offline socialization, then the managers of that community would have been stigmatized for moral failings and organizational fault and pressured to attempt impossible tasks to fix the issue. When everyone was doing enough posturing to acknowledge the problem, them it became easier for any community to state its own situation and advance the necessary discussion and collaboration. If there was a public imagining of a systemic problem before 2012, then it was low-level general incivility that might target any group which regularly faces conflicts, like LGBT+ people, individuals editing contentious political or religious topics, and women included but not particularly in need of special attention.
In April 2014 someone started the Wikipedia article “Gender bias on Wikipedia“, so that marks a point in time where there was enough published journalism on the topic to fill out a Wikipedia article. The fact of a Wikipedia article means that conversation in Wikipedia had been ongoing long enough for journalists and researchers to respond to that conversation, and then finally for someone to notice that enough publication had been done to create a Wikipedia article summarizing it.
In February 2015 the Wikimedia Foundation began what was to become a recurring grants program called “Inspire“, which piloted with a call for project proposals to increase engagement among women. Regardless of whatever else that program accomplished with actual grantmaking, that instance of the Inspire program led the Wikimedia community to have discussions about the extent to which outreach to women was necessary. Before that program, many people doubted it. After the program, with significant credit to Dorothy, word got around that harassment of women was a serious enough problem to merit an allocation of staffing and grants. I wish to emphasize that definitely at the time of that February 2015, repeatedly and in every way the prevailing thought and default assumption in any decision-making process was that the Wikimedia Foundation would leave the management of harassment to community volunteers and not have its own staff be involved in policy development, dispute resolution, best practice, or anything similar. Also, people who spoke for the Wikimedia Foundation publicly or in rumors had no intention to direct any significant amount of funding into this space. There were significant but quieter gestures before the Inspire campaign, but the Inspire campaign was a public and controversial demonstration that the WMF did care about this issue. Still, despite the Inspire campaign being a bold push, it was also cautious and reserved with limited staff time to make it as effective enough to meet community demand and limited funding to make the change. There was no fault with the campaign – the model started at an appropriate size and scope for a first iteration that needed to grow with time. The Inspire campaign was just entering an issue much bigger than it was designed to manage, and the intent of the campaign never was to address harassment anyway. That iteration of it was just supposed to promote participation by women by encouraging positive participation rather than imagining how to reduce negative activities.
I had been talking with Dorothy about harassment perhaps since she moved to New York in August 2013. My concern was about LGBT+ harassment and Dorothy knew more about harassment for women. I came to believe that since there are lots of women, and since they get the most harassment, then the best way to sort the harassment problem for LGBT+ and for any other minority group experiencing harassment would be to join the effort to solve the problem for women. When there would be a solution for harassment against women then that solution should also work for LGBT+ harassment or any other kind of harassment. I wrote in a March 2015 post “Logistics behind promoting and planning a long-term solution to online harassment with partners” about some of the challenges we were considering. The most significant takeaways that I had from early exploration of the topic was that in 2015, the philosophy at the Wikimedia Foundation was that addressing harassment with staff attention was something that was not likely to happen at any time in the forseeable future.
“Harassment” is something that should be defined, and it is challenging to summarizing the topic, but I mean the perception of being the target of hostility which deters a person from positive participation in the community. If harassment is managed on a scale of 1-10, with 1 being mild and 10 being severe, then I feel that the Wikimedia community addresses harassment on a level of 1-3 very well with volunteer infrastructure. The Wikimedia Foundation has always been available to address severe harassment at the 9-10 level, like death or legal threats. The challenge is always that mid-tier level 4-8 harassment. Problems addressing this harassment include cost, high failure rate in attempts to make the situation better, high liability in even getting involved, and high commitment required to serve all situations and languages globally if any cases are taken in any single situation in one region. I understand why the Wikimedia Foundation along with all other similar organizations found the idea of getting involved to be unthinkable; all options which anyone had presented would be costly, ineffective, and unsatisfying. However, around 2015 the way that Dorothy and I presented the issue in the IdeaLab was not as a problem which needed a solution, but instead as a problem which needed a commitment for long-term financial investment to address.
I think that no one has an easily comprehensible way to address harassment but there are some viable ideas in circulation that might, for example, target 1% of Wikipedia’s harassment cases and resolve them with 50% efficacy. Usually when I hear someone’s idea to solve the entire Wikipedia harassment problem, I would characterize it as a “1% of the problem, 50% of the time” solution, with all of these ideas being costly. Still, in the longer term Wikipedia has to do something and piloting some of these ideas and researching their efficacy could be a path to a solution. Something that Dorothy and I were allowed to say, but that Wikimedia Foundation staff were not allowed to say because of conflict of interest, was that we wanted the Wikimedia Foundation to create a documentation repository of complaints. Things developed and in December 2016 the WMF published some draft advice for Wikimedia volunteer community groups about documenting evidence of harassment as a way to support victims of harassment. I think the advice that someone should do documentation is sound, but I do not agree that crowdsourced volunteers can manage documentation in a way that does not compromise the safety of all involved and risk weird blame, accusations and making everything worse. In an office if there is a harassment complaint, the human resources department will not crowdsource a solution among the employees, but instead they document the issue in a professional way and keep the trust of everyone to limit access to the information they hold. It is quite beyond volunteer ability to do meaningful documentation.
Dorothy and I had the idea that the WMF would operate a complaint box. The idea was that anyone could report any complaint into that complaint box, and it would be documented safely with a time stamp. Access to any given complaint could be granted by the person who made it, but also, highly vetted data researchers with no interest in any particular case could have access to the complaints for the purpose of discovering fundamental unknown information about harassment. Basic questions which we cannot answer include how often people complain, why they complain, whether they ever find a resolution, how serious the issue is, and whether multiple people all have complaints about any single individual. We started developing the proposal for the Centralised harassment reporting and referral service in October 2014. This grant proposal – which was never to request funds for anyone or any organization in particular, and was always just a call for funding to be allocated to any reputable organization for crisis management – was turned down in April 2015 after review in that first Inspire campaign for women’s issues. The mood at the time and before this was that there would never WMF staff appointed to manage community social or personal crisis, and with the denial of the grant, there would be no funding for external support either. I understood the denial of the grant because our grant request was totally beyond the rules of the grant request process. That inspire grant series was more about funding small projects with small amounts of money, because again at the time – April 2015 – the Wikimedia Foundation was not ready to say that harassment was among the issues which were among its responsibilities to allocate funding to address. I am still glad that Dorothy and I applied. A consequence of our applying for this grant was that we raised the point in discussion that this was an issue which merited 100s of thousands of dollars in funding soon and in an organized way. Even if that money did not bring a solution, then at least it limited liability. Having a harassment crisis in the context of not investing in a solution is much different than having a harassment crisis when the problem is funded with staff, collaborations, research, and pilot projects. The point of the request was to establish an expectation that funding would go to crisis management and to that end it succeeded. Also we were forcing the conversation to compel the WMF to consider whether if there were a steady stream of 100s of thousands of dollars going to this issue, would that money go to an external organization, or would the WMF like to keep that money internally and use it to build out its empire? Framing the discussion as a question of who will be funded to address the problem made the conversation different than asking the WMF to hire someone internally to address the problem. As long as the conversation was about asking the WMF to take responsibility, they response was that the matter was out of scope. When the conversation was about which external university or nonprofit organization was going to get a community mandate to make a territorial claim in a high profile position of the Wikimedia community and get perpetual approval for disbursement of large amounts of funds managed by the WMF, then the conversation started trending toward the WMF feeling more able to provide some services by hiring staff in-house.
Dorothy and I contacted any likely organization would could find through search or referral to be the one to partner with the WMF to address harassment. We had wondered if there were a large project to address harassment, funded on the order of 300k, then who would manage the project? If a suitable organization existed with expertise in addressing harassment in online communities, then they should get the funds. We contacted various organizations. Some of them asked that we not publicly talk about them and Wikipedia, and with other organizations we thought better of asking them about a public conversation. In that time – 2014-2015 – I would not want anyone in the future to misunderstand that Wikipedia still had a poor reputation. There is a trend for Wikipedia’s reputation to be improving, but still Wikipedia is mostly reviled, and in the time that we were asking we were hardly taken seriously. Organizations which I felt were minor communicated with us with a perception that Wikipedia was a fad almost at its end and that there would be no merit in supporting its community base, even if paid with funding. I do not want to name names, but just imagine that Dorothy and I sent emails to about 10 likely big-enough organizations. Lots of them took us seriously for what we said, especially when we were offering them hundreds of thousands of dollars and an ongoing stream of funding. We had phone and video chats with many of them, some of them several times, and by their seeming lack of awareness of what we imagined to be the fundamental background of the issue, we came to believe that there would be no ideal organizational partner with a background in harassment management. For context, the professional backgrounds which we sought were in de-escalation, crisis hotline, diversity training, and online community management. My assertion is that in 2014-2015 major organizations operating in these fields had not yet seen value in applying their expertise to online communities. Even organizations which which had been recently founded specifically to address online harassment were far away from our expectations. Either they wished to force the application of offline techniques in irrelevant ways to online communities, or simply had what Wikipedians would call no meaningful experience with online communities, or they were just corporate minded and wanted to be highly funded perpetually to work mostly in secret without community consultation while periodically releasing products. I became satisfied that it would be easier to teach Wikipedia community leaders to learn harassment management than it would be to teach harassment management professionals what they needed to know about online communities.
Following our grant request, I felt like Dorothy and I had success as the WMF went contrary to everything which we had been told many times by many people before. The position of the WMF switched, and in fact, they did start investing in long term harassment management even with the understanding that in the short term this expensive investment would provide no relief.
In April 2014 Patrick accepted an appointment to the WMF’s community advocacy division. Patrick is a colleague and friend of both Dorothy and me, and we were glad to see him in the role. At the time there were no public plans to put more people into community advocacy, and not much news about what community advocacy would become, but Patrick said that first steps would be information collection, sorting what has already been done, doing surveys, and seeking community comment on various proposals. I am sure that Patrick was hired for his neutral and objective outlet and his good sense of personal boundaries and work-life balance, which are essential skills in the crisis business.
A strange thing about Patrick’s hire was that it was a reimagining of the concept of the “Community Advocacy” department, with Patrick given an unusual opportunity to express a lot of personal vision to influence both the direction of the department and the WMF position on harassment as a whole. When he was hired, the nature of the position it could have been perceived as a chump role just because as I said professionals in the field of harassment management mostly had no respect for Wikiepdia or even online communities. I had been watching the harassment situation play out as an epic and when Patrick placed himself into this unwanted role, I realized that he would impose the hegemony of the Cascadia ideal of virtue onto Wikipedia’s community values and by extension all human interaction in every context forever after. He was taking an undesirable humble job opening which also which had a greater worth and impact than almost anyone watching realized, and which was going to grow into something history making. He had a great skill set going into the job, and now he is a world authority in a hot field. I was delighted to see someone familiar and appropriate get this position.
For background context, in February 2012 the Wikimedia Foundation established a group of staff called “Community Advocacy”. I only know the work of this group as it faced the Wikimedia community, but I am no insider who knows what private projects it did. Between 2012 and 2014 it offered legal support to Wikipedia editors who were the target of legal threats for editing Wikipedia articles. At the time it was first imagined that legal threats were a major problem. They are, and every day all kinds of people from the Internet go to Wikipedia to post and email wild lawsuit threats. Dozens of these come in every day, but in contrast to the pre-Internet age when legal threats were more serious, nowadays the legal threats for wiki go into a pile for volunteers to address as with so many other wiki task queues. The volunteers who address these usually manage them with canned responses and laughs, but volunteers also need to be able to identify which legal threats they should send to the staff legal team at the WMF. Legal@wikimedia.org assists volunteers by issuing an escalated service tier of canned responses and laughs to people who write in requesting a more thorough response. In September 2012 a company targeted two Wikipedia volunteers with a lawsuit. The Wikimedia Foundation funded their legal counsel, and that led to a general commitment that if anyone is doing good on wiki and is the target of a lawsuit as a consequence, then the Wikimedia Foundation will fund that person’s legal counsel. One significance of that lawsuit was that it led to a recognition of the divergence of needs between legal issues and other sorts of issues. I would attribute the special needs of addressing that lawsuit to being one of the factors which made the Community Advocacy department more open to forking off to another focus, like harassment.
In 2013-14 there was a major conflict between the Wikimedia Foundation and the Wikimedia community over the rollout of the Media Viewer software. In summary, the sudden software changes without sufficient community input greatly disrupted the workflow of many Wikimedia community volunteers. The initial WMF response was to dismiss complaints or have any random staff person speak publicly about the issue instead of planning an institutional response, and hundreds of contributors consequently felt disenfranchised and mistreated. There hundreds of pages of text discussing that. Following that, the WMF made commitments to minimize harm in the future whatever else happened. One change that followed this was that in July 2014 the Community Advocacy team revised the page and added a statement saying, “We provide support on the rollout of major changes, such as legal policy updates and OTRS software upgrades.” The idea was that the group was transitioning from being a support service to Wikipedians from external threats, to also being a support service for protecting the community from harms that might come from the WMF’s own software changes. The team wanted to be available to hear any sort of serious problem that Wikimedia community members were experiencing. Software rollouts should never have been imagined as being approximately the same as addressing harassment or providing assistance to someone being sued, but again – it made sense at the time to combine all these things, and this particular issue was the primary cause for separating software rollouts from harassment.
In January 2016 the Community Advocacy team changed its name and their to “Support and Safety” and also changed the mission of the team. This was a consequence of Patrick’s April 2014 hire into this role as harassment person. The team hired other people, and different people have different roles, and it is not as if Patrick was a mastermind of the process, but he had a major role and influence as the designated starting point of reform. In the context of legal being managed in a different way now, and software rollouts being managed in a different way, that created an opening for “community advocacy” to differentiate itself as “harassment management”.
AT WikiConference North America 2016 Patrick and I talked about community guidelines he had been developing to outline the two big concepts – “Dealing with Online Harassment” and “Keeping events safe“. I told him that I would share these with Wikimedia New York City and that our chapter would be the first adopters of these policies, because we already had been trying to draft our own. Now with these well-prepared drafts to start the conversation, whenever we as a Wikimedia community chapter have a problem and someone pressures us to do something about it, we can point to these guides and explain our position. On Wednesday 15 February 2017 after many small group and interpersonal discussions we introduced the drafts to our public WikiWednesday meeting. That does not mean that everything is sorted, and neither our chapter board accepting the guides nor acceptance at any public meeting would mean that the last word has been said, but now at least with these guides we will never have to begun a discussion from the beginning. The problem remains that easy problems are easy to resolve, and hard problems are easy to pass on, but those mid-range problems are going to continue to be a challenge and guides like this are an emotional, social, and organizational support that make a big impact on how volunteers related to each other and feel about themselves. I am very happy with the pace and output of the Wikimedia Foundation’s investment in this space.