Democratic innovations have been widely presented by both academics and practitioners as a potential remedy to the crisis of representative democracy. Many argue that deliberation should play a pivotal role in these innovations, fostering greater citizen participation and political influence. However, it remains unclear how digitalization affects the quality of deliberation—whether digital democratic innovations (DDIs) undermine or enhance deliberation. This paper takes an inductive approach in political theory to critically examine three features of online deliberation that matter for deliberative democracy: (...) scale, transparency, and the facilitation of equality. It argues that the enhancement of equality should be given special attention when designing and implementing deliberative DDIs. Equality is a basic democratic value that is crucial for the intrinsic quality of deliberation. One way of enhancing equality through digital deliberative practices is via AI facilitation, which can assist human facilitators and help mitigate power dynamics, often present in non-digital settings. (shrink)

It is well-established that algorithms can be instruments of injustice. It is less frequently discussed, however, how current modes of AI deployment often make the very discovery of injustice difficult, if not impossible. In this article, we focus on the effects of algorithmic profiling on epistemic agency. We show how algorithmic profiling can give rise to epistemic injustice through the depletion of epistemic resources that are needed to interpret and evaluate certain experiences. By doing so, we not only demonstrate how (...) the philosophical conceptual framework of epistemic injustice can help pinpoint potential, systematic harms from algorithmic profiling, but we also identify a novel source of hermeneutical injustice that to date has received little attention in the relevant literature: epistemic fragmentation. Epistemic fragmentation is a structural characteristic of algorithmically-mediated environments that isolate individuals, making it more difficult to develop, uptake and apply new epistemic resources. This, in turn, can impede the identification and conceptualisation of emerging harms in these environments. We trace the occurrence of hermeneutical injustice back to the fragmentation of the epistemic experiences of individuals, who are left more vulnerable by the inability to share, compare, and learn from shared experiences. (shrink)

Unsupervised learning algorithms are widely used for many important statistical tasks with numerous applications in science and industry. Yet despite their prevalence, they have attracted remarkably little philosophical scrutiny to date. This stands in stark contrast to supervised and reinforcement learning algorithms, which have been widely studied and critically evaluated, often with an emphasis on ethical concerns. In this article, I analyze three canonical unsupervised learning problems: clustering, abstraction, and generative modeling. I argue that these methods raise unique epistemological and (...) ontological questions, providing data-driven tools for discovering natural kinds and distinguishing essence from contingency. This analysis goes some way toward filling the lacuna in contemporary philosophical discourse on unsupervised learning, as well as bringing conceptual unity to a heterogeneous field more often described by what it isnot(i.e., supervised or reinforcement learning) than by what itis. I submit that unsupervised learning is not just a legitimate subject of philosophical inquiry but perhaps the most fundamental branch of all AI. However, an uncritical overreliance on unsupervised methods poses major epistemic and ethical risks. I conclude by advocating for a pragmatic, error-statistical approach that embraces the opportunities and mitigates the challenges posed by this powerful class of algorithms. (shrink)

Over recent years, the stakes and complexity of online content moderation have been steadily raised, swelling from concerns about personal conflict in smaller communities to worries about effects on public life and democracy. Because of the massive growth in online expressions, automated tools based on machine learning are increasingly used to moderate speech. While ‘design-based governance’ through complex algorithmic techniques has come under intense scrutiny, critical research covering algorithmic content moderation is still rare. To add to our understanding of concrete (...) instances of machine moderation, this article examines Perspective API, a system for the automated detection of ‘toxicity’ developed and run by the Google unit Jigsaw that can be used by websites to help moderate their forums and comment sections. The article proceeds in four steps. First, we present our methodological strategy and the empirical materials we were able to draw on, including interviews, documentation, and GitHub repositories. We then summarize our findings along five axes to identify the various threads Perspective API brings together to deliver a working product. The third section discusses two conflicting organizational logics within the project, paying attention to both critique and what can be learned from the specific case at hand. We conclude by arguing that the opposition between ‘human’ and ‘machine’ in speech moderation obscures the many ways these two come together in concrete systems, and suggest that the way forward requires proactive engagement with the design of technologies as well as the institutions they are embedded in. (shrink)

Technologies of “artificial intelligence” and machine learning are increasingly presented as solutions to key problems of our societies. Companies are developing, investing in, and deploying machine learning applications at scale in order to filter and organize content, mediate transactions, and make sense of massive sets of data. At the same time, social and legal expectations are ambiguous, and the technical challenges are substantial. This is the introductory article to a special theme that addresses this turn to AI as a technical, (...) discursive and political phenomena. The opening article contextualizes this theme by unfolding this multi-layered nature of the turn to AI. It argues that, whereas public and economic discourses position the widespread deployment of AI and automation in the governance of digital communication as a technical turn with a narrative of revolutionary breakthrough-moments and of technological progress, this development is at least similarly dependent on a parallel discursive and political turn to AI. The article positions the current turn to AI in the longstanding motif of the “technological fix” in the relationship between technology and society, and identifies a discursive turn to responsibility in platform governance as a key driver for AI and automation. In addition, a political turn to more demanding liability rules for platforms further incentivizes platforms to automatically screen their content for possibly infringing or violating content, and position AI as a solution to complex social problems. (shrink)

The negative impacts of social media have given rise to philosophical questions around whether social media companies have the authority to regulate user-generated content on their platforms. The most popular justification for that authority is to appeal to private ownership rights. Social media companies own their platforms, and their ownership comes with various rights that ground their authority to moderate user-generated content on their platforms. However, we argue that ownership rights can be limited when their exercise results in significant harms (...) to others or the perpetration of injustices. We outline some of the substantive harms that social media platforms inflict through their practices of content moderation and some of the procedural injustices that arise through their arbitrary application of community guidelines. This provides a normative basis for calls to better regulate user-generated content on social media platforms. We conclude by considering some of the political and legal implications of our argument. (shrink)

AI seems like the perfect response to the growing challenges of content moderation on social media platforms: the immense scale of the data, the relentlessness of the violations, and the need for human judgments without wanting humans to have to make them. The push toward automated content moderation is often justified as a necessary response to the scale: the enormity of social media platforms like Facebook and YouTube stands as the reason why AI approaches are desirable, even inevitable. But even (...) if we could effectively automate content moderation, it is not clear that we should. (shrink)

This article explores the cooperation of government and the private sector to tackle the ethical dimension of artificial intelligence. The argument draws on the institutionalist approach in philosophy and business ethics defending a ‘division of moral labor’ between governments and the private sector. The goal and main contribution of this article is to explain how this approach can provide ethical guidelines to the AI industry and to highlight the limits of self-regulation. In what follows, I discuss three institutionalist claims. First, (...) principles of AI ethics should be validated through legitimate democratic processes. Second, compliance with these principles should be secured in a stable way. Third, their implementation in practice should be as efficient as possible. If we accept these claims, there are good reasons to conclude that, in many cases, governments implementing hard regulation are in principle the best instruments to secure an ethical development of AI systems. Where adequate regulation exists, firms should respect the law. But when regulation does not yet exist, helping governments build adequate regulation should be businesses’ ethical priority, not self-regulation. (shrink)

The key role played by online platforms in the neo-intermediation of the public debate requires a review of current tools for mapping the digital information ecosystem, highlighting the political nature of such an analysis: Starting from a synoptic overview of the main models of platform governance, we try to understand whether the ongoing European shift towards the Limited Government Regulation (LGR) model will be able to counterbalance the “systemic opinion power” of the giant platforms and restore the “health” of the (...) digital information ecosystem. A close analysis of the European Digital Services Act (DSA) has highlighted some limitations in achieving its goals, because of the features of the LGR on the one hand, and the disruptive features of the algorithmic neo-intermediation phenomenon on the other. Thus, we suggest a tripartite regulatory model, that can be defined as “neo-editorial accountability.” However, increasing users’ critical algorithmic awareness is an essential prerequisite for implementing the suggested template, and mitigating an outstanding effect of the LGR model: the normalization of the ideological assumptions underlying informational capitalism. (shrink)

Effective content moderation by social platforms is both important and difficult; numerous issues arise from the volume of information, the culturally sensitive and contextual nature of that information, and the nuances of human communication. Attempting to scale moderation, social platforms are increasingly adopting automated approaches to suppressing communications that they deem undesirable. However, this brings its own concerns. This paper examines the structural effects of algorithmic censorship by social platforms to assist in developing a fuller understanding of the risks of (...) such approaches to content moderation. This analysis shows that algorithmic censorship is distinctive for two reasons: in potentially bringing all communications carried out on social platforms within reach and in potentially allowing those platforms to take a more active, interventionist approach to moderating those communications. Consequently, algorithmic censorship could allow social platforms to exercise an unprecedented degree of control over both public and private communications. Moreover, commercial priorities would be inserted further into the everyday communications of billions of people. Due to the dominance of the web by a few social platforms, this may be difficult or impractical to escape for many people, although opportunities for resistance do exist. (shrink)

Social media platforms have been rapidly increasing the number of informational labels they are appending to user-generated content in order to indicate the disputed nature of messages or to provide context. The rise of this practice constitutes an important new chapter in social media governance, as companies are often choosing this new “middle way” between a laissez-faire approach and more drastic remedies such as removing or downranking content. Yet information labeling as a practice has, thus far, been mostly tactical, reactive, (...) and without strategic underpinnings. In this paper, we argue against defining success as merely the curbing of misinformation spread. The key to thinking about labeling strategically is to consider it from an epistemic perspective and to take as a starting point the “social” dimension of online social networks. The strategy we articulate emphasizes how the moderation system needs to improve the epistemic position and relationships of platform users — i.e., their ability to make good judgements about the sources and quality of the information with which they interact on the platform — while also appropriately respecting sources, seekers, and subjects of information. A systematic and normatively grounded approach can improve content moderation efforts by providing clearer accounts of what the goals are, how success should be defined and measured, and where ethical considerations should be taken into consideration. We consider implications for the policies of social media companies, propose new potential metrics for success, and review research and innovation agendas in this regard. (shrink)