Actors’ views : Willie Robert

1.   Can you tell us about the mission of the Wikimedia France association and your main challenges today?

The Wikimedia Foundation is an American charity, in charge of the technical infrastructure (hosting, bandwidth, software) of the free encyclopedia Wikipedia and its sister projects (Wiktionary, Wikidata, Wikisource, Wikimedia Commons, etc.). However, it does not directly intervene on the actual content (encyclopedic articles, pictures, etc.). The whole content is self-managed by communities of volunteers, i.e. virtually any Internet user. At the local level, associations such as Wikimédia France aim to support free knowledge and promote projects such as Wikipedia. These associations are legally independent from the Wikimedia Foundation and do not intervene on the contents either. The actions of Wikimédia France include public awareness campaigns, organizing and/or participating in national and international conventions and conferences, and supporting the community of contributors to Wikimedia projects.

For example, Wikimedia France mobilizes its expertise to help political, educational, and cultural institutions to understand how Wikimedia projects work and how to take part in them, by contributing articles or releasing content. This is a large part of the association’s activity. Another current challenge is to respond to the recent proliferation of national and European legislation aimed at regulating the digital world, with the upcoming Digital Service Act, which will undoubtedly be the most important European regulation in this area for the next 20 years. We are here to make institutions understand that the Internet is not only about a few very large multinational companies, and that legislation must not “throw out the Wikimedia baby with the GAFAM bath water”. Unfortunately, it is quite difficult in 2020 to be heard as a free, voluntary and open source-based model, whose content can be reused under free license, which relies on donations and without any advertising or exploitation of personal data… even though it is being used by tens of millions of people every day.

2.   What current and future challenges is Wikipedia facing (or could be facing), in relation to its community-based content moderation model?

Wikipedia’s moderation system is unusual because it is based on citizens and volunteers. This means that anyone, at any time, can moderate its content. Any Wikipedia reader is only a few clicks away to edit the content of an article, without even having to create an account or log in since it is not mandatory to participate in Wikipedia. This is one of the strengths of our model: this openness which allows reactivity. Currently, around fifty to sixty thousand people contribute to the French version of Wikipedia every month, fifteen to twenty thousand of whom have created an account. In this regard, we have identified two main challenges: maintaining a high level of participation and increasing the community’s diversity.

Regarding high participation, some pitfalls to be avoided are technical difficulty (for example, it is less convenient to write articles on a mobile phone than on a computer, although it is well known that uses tend to shift towards smartphones) and neglecting inclusiveness (having a broader, more welcoming community, doing better at managing the inevitable conflicts that occur just like in any social group, etc.). This is part of the work conducted by the Wikimedia Foundation, which manages technical aspects but also takes initiatives on welcoming community members, fighting online harassment, etc., in liaison with local communities and associations.

The second challenge is the diversity of the community. More diversity, of course, leads to higher overall participation, but it goes beyond that. If we want to fight against the proliferation of all kinds of hateful or discriminatory content, the community must be able to identify it even when it is expressed in a low-key or roundabout way. Making sure that the people who contribute to Wikipedia are as diverse as possible is therefore necessary to guarantee a balanced and sensitive moderation of all issues. Then, there is also the question of invisibility: removing this or that insult is micro-moderation. This is important, but it is not the most complicated issue and above all, it is not enough. It is also necessary to engage in macro-moderation, i.e. to ensure that the entire « discourse » expressed through the millions of Wikipedia articles is not in itself discriminatory. It is indeed possible that one theme or point of view is overemphasized while another is completely invisible, or on the contrary that a given subject is completely left aside, and therefore hardly moderated, because the community lacks people who are interested in it. Representation biases are among the first vectors of discrimination and toxic discourse. Wikipedia is affected by the same biases as society, of which it is a reflection: in order to reduce these biases and, at the same time, improve our moderation model, we need to increase the participation of women, of French-speaking people from all over the world (especially Africa), to attract more academics, more people from socially underprivileged backgrounds, more young people, more elderly people, and so on. In short, we need more diversity.

3.   What do you think of the emerging debates over the end of online anonymity? Would it be an effective way to counter the spread of toxic content online?

Wikimedia France is particularly careful on this issue because challenging anonymity would expose the encyclopedia’s operation to significant risk. It would be a disaster for the elaboration of articles, for the moderation of contents and for the community. Beyond these matters, we are also very concerned about freedom of expression in general, especially because the problem is usually poorly framed. Anonymity and pseudonymity are frequently confused, and the underlying technical realities are most often ignored. While Internet communications often highlight a deeply self-centered aspect of social life, Wikipedia’s model for content production is based on putting egos aside. The basic principles are that people who write in Wikipedia never write on their own behalf, but transcribe the views expressed by third-party sources, citing those sources whenever there is a disagreement or to prevent any disagreement. This is key to building consensus, through dialogue and confrontation of sources. And, as a matter of fact, this mandatory search for consensus results in a very limited spread of toxic content on Wikipedia, by design. Putting an end to anonymity would not improve anything on Wikipedia, quite the contrary.

Regarding the moderation of online toxic content more specifically, volunteers who are identifiable and are involved in moderation are also particularly exposed. Indeed, many Internet users do not accept being blocked, banned or having their work erased. This has sometimes gone as far as online harassment or employers being contacted because of a dispute. Being able to identify a moderator could be an important and dangerous form of pressure, which often harms the weakest members of society. Having everyone expressing themselves under their civil identity, as some politicians suggest, is in fact quite simply pernicious.

Generally speaking, the possibility for an individual to hide his or her identity is a way of protecting themselves: just as some of the authors of Diderot and d’Alembert’s encyclopedia, many Wikipedians chose not to reveal their identity for safety reasons. Indeed, even today, in some parts of the world, it can be dangerous to write about the head of State, political opposition, religion etc. When we, in France, talk about “anonymity”, do we realize that our digital space is global, and that what we read may have been written by a French-speaking individual living in a dictatorial regime? Do we also realize that, even in France, it may still be difficult for one to express himself or herself as a member of the LGBTQ+ community?

Do we really want to let people choose between coming out or keeping silent as the sole alternative? Are we thinking about minors? Will they also have to reveal their identity to the whole web, and lose their right to have their youthful mistakes forgotten? Indeed, some malicious people do hide their identity to spread toxic content. But the fact that the cure is worse than the disease is hardly emphasized, and this is worrying. Not to mention that the effectiveness of such measures remains to be proven: whatever « solution » would be chosen, we also know that the people who are most motivated and best equipped to circumvent this type of legislation are often the most malicious people.

Finally, there is also some confusion as to who is being targeted by the debates related to the end of anonymity on the « Internet ». We generally talk about « social networks », but we should not be mistaken: if a law were to be passed, it would have a much broader reach than just the two or three social networks generally mentioned under some vague terms (« platform », « content operator », etc.). The same semantic shortcuts were used during the debates around the law against information manipulation and the law against hateful content. Because of these shortcuts, Wikipedia falls within the scope of the first law and would have been equally concerned by the second, even though it is not identified by anyone as being a vector for the problems we are trying to solve.

Rather than conveying simplistic ideas, we are campaigning for education (of Internet users and political leaders), for more coordination between the different actors involved and for more means to be allocated to the justice system in order to act while respecting fundamental freedoms.