Early Impressions: How Europe’s Proposed Digital Services Act Can Preserve Wikimedia, or Let it Get Constantly Trolled

The Wikimedia Foundation’s take on the DSA: Cautious optimism, but also concerns about empowering bad-faith actors

Wikimedia Foundation Policy
4 min readDec 23, 2020
The image shows railroad crossings in Buenos Aires in 1908. It was published in the Electric Railway Journal.
Railroad Crossings (Electric Railway Journal, 1908)

The European Commission recently released its proposal for the Digital Services Act (DSA), a law that will change the legal underpinnings of online life in Europe, and, by extension, the world. One of the main components of the proposal creates a framework of obligations for online hosts — a group which includes the Wikimedia Foundation in its role as the host of Wikipedia.

The current law on the liability of hosts, governed by Article 14 of the e-Commerce Directive, says that online hosts aren’t liable for what their users post if they don’t know about any illegal activity, and if they act upon illegal activity once they know about it. Article 15, meanwhile, says that a host can’t be legally required to monitor its services, hunting for any potentially illegal activity by its users.

There’s a lot to analyze and consider in the DSA proposal, but we wanted to share a few early impressions. First of all, we’re glad to see that the DSA preserves these provisions of the e-Commerce Directive, which ensure that the Foundation can continue hosting the knowledge of countless editors and contributors. Unique, user-driven platforms like Wikipedia thrive under strong intermediary liability protections, and we are happy to see this recognition from the Commission. In addition, there are lots of new provisions in the DSA as well, intended to encourage more effective and responsive content moderation. While some of these improve transparency, such as making it easier for people to understand why they see a certain piece of information, and are intended to promote fundamental rights, there are also others that, if applied poorly, could actually make some hostile aspects of the internet worse.

In particular, our initial concerns center on two things. The first is in Article 12, which says that an online service has to disclose in its terms of service the rules and the tools of its content moderation. While we agree that terms of service should be as clear and transparent as possible, our concern lies with language like that in Article 12.2, which says that online services must apply and enforce their terms of service “in a diligent, objective, and proportionate manner.” That’s an ideal goal, but we worry that “diligent, objective and proportionate” can mean very different things depending on who you ask, and that community-governed platforms would be hurt by unclear standards and a lack of discretion. Terms of use (like the Foundation’s Terms, or even the Universal Code of Conduct) frequently include provisions prohibiting clearly harmful but often hard-to-define and even platform-specific things like harassment, disruptive behavior, or trolling. At what point would a regulator or a litigious user think that a certain volume of trolling meant that a service wasn’t being “diligent” in enforcing its “don’t troll other users” rule? Or what happens when someone whose posts are moderated, or who thinks someone else’s behavior should be moderated, decides that the moderators aren’t being “objective?” These situations certainly happen often enough, but usually don’t give rise to legal disputes. Under the proposed rule, we are concerned that the resulting uncertainty about what “diligent, objective, and proportionate” moderation should be would lead disgruntled users to bring such costly cases while the world waits for more definitive and uniform guidance from the European Court of Justice. Keep in mind this is taking place in the context of an information sphere currently struggling with countless motivated and well-funded bad-faith arguments, disinformation, and conspiracy theories. And the number of disgruntled users is only going to grow.

The other concern we want to raise comes up in Article 14, which says that an online provider will be presumed to know about illegal content — and thus be liable for it — once it gets a notice from anyone that that illegal content exists. There’s a number of different ways that ambiguities in this section can create problems, including potentially contradicting the prohibition on general monitoring obligations. For example, if the Foundation got a notice from someone alleging they had been defamed on one article, what would the Foundation be responsible for, if the alleged defamation was referenced in or spread across multiple articles, or talk pages, that the user may not have specified? There must be significantly more clarity around this provision if it is going to operate as intended and not pose an undue burden on platforms.

Finally, we want to make sure that the particular structure, mission, operation, and community self-governance of Wikimedia projects and other participatory platforms are accounted for in this piece of legislation that was likely designed with very different kinds of platforms in mind. We still see some gaps and omissions in the Commission’s proposal and look forward to collaborating with colleagues and members of the Wikimedia movement in Europe (with a particular shout-out to the tireless work of the Free Knowledge Advocacy Group EU) to work with lawmakers to ensure that the law can support and foster the kind of free, open, collaborative, and collegial space that is the best of the Wikimedia movement.

Sherwin Siy, Lead Public Policy Manager, Wikimedia Foundation
Allison Davenport, Senior Public Policy Counsel, Wikimedia Foundation
Jan Gerlach, Lead Public Policy Manager, Wikimedia Foundation