Tor (The Onion Router) is a digital infrastructure that has been persistently taken to task by the media for its evilness. While it makes anonymous Internet navigation possible, Tor also grants access to the so-called darknet, where illegal goods (such as drugs, weapons, child pornography, and counterfeit market items) and terrorist communications circulate widely. The moral panic in response to the malicious content circulating on the darknet is clearly good grounds for addressing the evil of this network. By contrast, I adopt an infrastructuring approach (see Star and Bowker 2002) to Tor: rather than taking into account the evil content flowing through the network, my interest is in the evil consequences that radiate out from within Tor as digital infrastructure, from the conflicts and controversies regarding the management of Tor to its architecture and political assets. In doing so, my aim is thus, as Andrew Barry has recently put it, to bring infrastructure’s “cracks and fissures, and the politics of their present and future condition” to the foreground.

The Tor network is a worldwide infrastructure made up of around seven thousand machines (or nodes) running dedicated software and protocols. Not only does the network allow evil content to circulate, but it also enables democratic and progressive activities to take place on the Internet (whistleblowing, political expression in authoritarian countries, browsing with greater privacy, etc.). Using a specific browser, users connect to Tor to surf the Internet anonymously and also to access hidden services, primarily anonymous websites only available through Tor’s network. While this infrastructure is peer-to-peer based, the working group behind it has been managed since 2004 by the centralized, U.S.-based, volunteer-operated organization called the Tor Project. The Tor Project itself is the work of the Electronic Frontier Foundation, an NGO that has turned the network into a flagship project aimed at defending privacy and digital rights. In recent years Tor has been connecting one to two million users on a daily basis, and the cracks and fissures related to the management of its infrastructure went largely unnoticed until very recently, when a series of controversies—which soon turned into heated clashes—revealed how a virtuous infrastructure aimed at defending rights can start to look evil when its inner workings and disagreements are revealed.

The Tor infrastructure’s first strains of evil date back to its invention—a sort of ancestral sin—as the network’s protocols were originally developed between 1996 and 2002 by the Naval Research Laboratory (NRL) in order to allow covert U.S. military personnel to communicate without being tracked and recognized. In 2004, the NRL decided to release these protocols publicly—if spies were the only individuals navigating anonymously, after all, it would have been obvious that these users were members of the military. Since then, controversy has been fueled by the fact that, after more than a decade, the links between Tor and the U.S. military have not been definitively cut. The website Pando, for example, has published articles claiming that the bulk of Tor funding is still mainly coming from U.S. governmental agencies. The emergence of other opaque links with U.S. agencies (e.g., the short-term hiring of an ex-CIA agent) has only added to the controversy. Indeed, it is not surprising that these shadowy connections would raise further suspicions for the many Internet activists for whom the National Security Agency is the greatest evil on the digital earth.

Controversy over Tor’s management and control exploded in 2016 when accusations of internal sexual harassment were made against a key Tor member and public figure, Jacob Appelbaum, who was pressured to leave the project along with other members linked to him. Meanwhile, a full turnover of the managing board marked a sharp discontinuity in project management. A number of high-profile contributors, including Lucky Green and Marie Gutbub, left the organization, alleging ethical concerns and triggering public debate. In the wake of these events, the Tor Project’s staff launched a general strike on September 1, 2016, as well as a project fork, with some contributors vowing to work on a parallel version. The reaction to these moves has been mixed, in part because of concerns over their potential impact on Tor users living under oppressive regimes.

Hence, in less than a year, controversies over Tor infrastructure management and control turned into a public drama, which damaged the network both in terms of public support among activists and lay users and with regard to the infrastructure’s technical performance. At the height of the controversy in late August, the number of bridges (servers allowing users to navigate anonymously on the Internet) fell back to 2013 levels.

The recent events surrounding the Tor network bring to light the specific cracks and fissures of Tor as a digital infrastructure, a sort of moral fracture that has emerged from the inside in the context of relationships and shared visions about the political and social aims of the project. A specific feature of Tor is that it represents a hybrid infrastructure: it is partly a U.S.-based institutional project and is simultaneously globally dispersed; an anonymous and anarchy-driven peer-to-peer architecture involving a considerably composite set of actors, from government-funded institutional NGOs to informal networks of hackers and radical activists plus, of course, a large base of dispersed lay users and project supporters. This hybrid configuration is clearly part of the conditions for the development of the controversies that surround Tor, which call into question the ethics, motives, and desires driving the people who cooperate in building and maintaining the infrastructure. The problems hidden (or suspected to be hidden) behind the visible surface of the project open a space of reflection on how the smell of evil can affect the ability to go on fostering the original good aims of an infrastructure.