Dear all Our AoIR pre-conference: "Algorithms and the future of agency" will take place online Monday October 26th at 17.00 UTC. Descriptions and program below. Organizers are: Jakob Linaa Jensen, Stefania Milan, Mark Andrejevic, Dylan Wittkover, Andrew Iliadis, Emiliano Treré, Aline Shakti Franzke. Sign up here: https://forms.gle/sS64W8MXX8ej3AsQ8 Zoomlink to be provided to participants the day before Idea: As more and more daily and social life is mediated through platforms, algorithms play an increasing role determining our information and capacity for action. For instance, the increasing use of automated decision-making, by insurance companies, hospitals, police and government in general raises important ethical and philosophical questions on the future of agency. First, in algorithmic procedures and decision-making, it is difficult to ascribe agency and thereby responsibility. Second, if our independent agency, also known as the free will, is severely limited by algorithmic decision-making and algorithmically defined horizons of information (for instance on Facebook and Google), it bears important and far-reaching consequences for liberal democracy as we know it. Responsibility, accountability and agency are fundamental prerequisites of democracy. If it is impossible to identify power and keep certain actors accountable, democracy and individual rights are at stake. In this pre-conference, we invite participants to explore and discuss such issues. To initiate and facilitate the discussions, a range of scholars with different background and interests, but all interested and engaged in the topic, will provide igniting talks discussing aspects of algorithmic agency. The first four talks sketch the problems from political, democratic, semantic and ethical perspectives. The last three talks sketch possible solutions, both conceptual and practical. We hope by the talk to inspire discussions based on specific experiences with algorithmic agency and to provoke suggestions for possible solutions in order to save agency and, ultimately, life and democracy as we know it in an increasingly algorithmic world. Program In the first half we will discuss the problems of algorithmic agendy and the first four organizers will give igniting talks (app. 10 minutes each followed by questions and comments from the audience). In the second half, we will discuss possible solutions to the problems, based on the igniting talks by the last three organisers. At the end of the session we will try to draw a map of these issues and questions that might form a backbone of a coming joint publication for interested participants. PART I – The problems of algorithmic agency Political agency in the age of AI: The case of facial recognition tech Stefania Milan How does artificial intelligence, and facial recognition technology in particular, alter the exercise of political agency? Taking the perspective of ordinary citizens, the contribution will touch upon two key aspects of democratic life, namely political participation and the relation between the state and its citizens. Briefly teasing out the possibilities and the perils for active citizenship entrenched in the expansion of AI into society, this contribution will raise questions on the future of democracy in relation to a technology which will have unparalleled consequences on human life. Semantic Agency: Graphing Knowledge on Web Infrastructures Andrew Iliadis Agency is tied to the power to name, classify, categorize, and organize. In today’s web infrastructures, knowledge about entities and relationships that exist in the world is codified into semantic web code/syntax that undergird countless information products, from Google knowledge panels to the Wikidata that feeds into virtual assistants. Such graphing of knowledge impacts areas as diverse as news consumption, fact checking, advertising, and logistics. This contribution draws on recent theoretical work on the philosophy and politics of web semantics (Monea 2016; Hui 2016; Thomas 2018) while also drawing on empirical research and case studies that focus on the work of platform companies and organizations who build and maintain semantic infrastructures on the internet. Algorithmic agency and the future of democracy Jakob Linaa Jensen Agency is closely related to the power we willingly or unwillingly impose towards ourselves in the platform economy, for instance when we sacrifice privacy for convenience. But such power is blurred and difficult to identify. It is a fundamental democratic principle that actors can be held responsible for their actions. If this is not possible, it has far-reaching consequences for liberal democracy as we know it. Ethical challenges of algorithmic agency Aline Shakti Franzke Agency has been a key term of ethics since Aristotle. In order to hold something accountable classical definitions have ever since focused on intentionality and mental representation. These human centered definitions, however, are stressed when applied to algorithms. Therefore, new ethical challenges occur along three lines a) meta-ethical concerns b) the agency of algorithms themselves and c) the implications of algorithms for human agency. The talk will map existing ethical debates along these axes. PART II – Possible solutions
From autonomy to respectful interdependence Dylan Wittkower
Postphenomenological analysis is less frequently used by Internet researchers, and offers distinctive contributions to an understanding of the future of agency. Postphenomenology avoids the reductive extremes of technological neutrality or technological determinism by offering descriptions of the lifeworld as constituted by hybridized human-technology agents (in embodiment technics) that interact with hybridized technology-world objects (in hermeneutic technics). I argue that this postphenomenological description of technological mediation encourages us to seek the future of agency not in autonomy or control, but in in respectful interdependence as articulated by feminist ethics of care. How “smart” technologies bypass subjective agency Mark Andrejevic Recent Foucault-inflected work on "environmentality" explores the ways in which environmental level monitoring and modulation (as envisioned in some versions of the "smart" city, for example) can operate as a strategy for control that bypasses subjective agency. The so-called libertarian paternalism of "nudge" economics, for example, might be described as a popularized form of environmentality that manipulates behavior below the threshold of reflexive consciousness. Such forms of manipulation come to rely on increasingly comprehensive forms of data capture designed to generate correlational predictions about future behavior under modulated circumstances. These developments emerge alongside ongoing attempts to bypass mediation by tapping straight into the human brain (as in the case, for example, of Elon Musk's Neuralink). I'll talk about possible alternatives and challenges to such developments. Resisting algorithms Emiliano Treré & Tiziano Bonini In this talk, we propose a conceptual framework to map the practices that citizens perform to shape, interfere and resist algorithms. Mobilizing the concept of the moral economy, we first situate the various forms of algorithmic agency along a continuum shaped by the two competing moral economies of the users and the platforms. Secondly, we bring into the equation the type of power held by those who enact specific forms of algorithmic resistance, adding the strategic vs. tactical poles. Finally, we discuss our framework drawing on examples including data obfuscation, streaming frauds, audience boosting, rating improvement, profile optimization, visibility enhancing, engagement groups, etc.