Academic journal article Yale Journal of Law & Technology

Tackling the Algorithmic Control Crisis: The Technical, Legal, and Ethical Challenges of Research into Algorithmic Agents

Academic journal article Yale Journal of Law & Technology

Tackling the Algorithmic Control Crisis: The Technical, Legal, and Ethical Challenges of Research into Algorithmic Agents

Article excerpt

TABLE OF CONTENTS  I. INTRODUCTION II. THE NEED FOR RESEARCH ON ALGORITHMIC AGENTS III. THE STATE OF RESEARCH INTO ALGORITHMIC AGENTS: THE AUDIT APPROACH IN PERSONALIZED MEDIA RESEARCH, AND BEYOND   A. The Algorithm Audit Approach   B. Beyond the Audit of Algorithms: Inquiry into the      User IV. THE CHALLENGES OF DESIGNING AN ALTERNATIVE: ROBIN--OUR MONITORING TOOL FOR ALGORITHMICALLY PERSONALIZED DIGITAL MEDIA   A. The Technical Design of Robin   B. Robin: The Challenges of Design and Implementation V. BALANCING INDIVIDUAL RIGHTS AND SOCIETAL NEEDS IN RESEARCH INTO ALGORITHMIC AGENTS: SOME LESSONS LEARNED   A. Designing Ethical Research: Transparency   B. Deploying the Principles in Practice   C. Creating Organizational Safeguards VI. CONCLUSION: TOWARDS A VISION ON RESPONSIBLE BIG DATA RESEARCH VII. ANNEX 1: JUSTIFICATION FORM FOR WEBSITE WHITELISTING VIII. ANNEX 2: WHITELISTED WEBSITE CATEGORIES   A. Sample Justification 

I. INTRODUCTION

Algorithmic agents permeate every instant of our online existence. As Artificial Intelligence research makes steady advances, as sensors proliferate, and more and more data are being accumulated and shared on data markets, the effectiveness of algorithmic recommendations grows while the costs of personalization drop. Consequently, many wonder if there will still be a space in the future where we remain insulated from direct or indirect exposure to algorithmic agents. In the age of ubiquitous algorithmic agents, will there be still spaces where we are not subjected to A/B tests, tailored advertising, price discrimination, and content recommendations? Will there be spaces in the future which are not controlled, one way or another, by algorithmic agents, and where technology is a neutral arbiter of rather than an active agent in our interactions in and with our environment?

Technology, as always, is deployed before society had the opportunity to come to terms with it. The lack of insight leads to a sense of lost control, drawing anxious responses. (10) Many see algorithmic agents as black boxes, (11) or rather, as black holes, which utilize all available information and grow ever powerful, but still remain invisible to human perception. Just like astrophysicists, scholars of algorithmic agents try to evaluate circumstantial evidence to understand how algorithmic agents operate, but unlike natural scientists, the researchers who study the sociological, political, economic, anthropological, ethical, and legal aspects of algorithmic black boxes regard their object of study as anything but natural or value-neutral. Algorithmic agents, just like any other technology, are embedded in the existing economic, social, and political conditions. They reflect our implicit and explicit hopes and fears, ambitions and shortcomings, and the social conditions in which they are created and used. (12) This means that, despite many techno-optimistic accounts, (13) the discourse on algorithms also reflects the fears and speculations on the adverse effects of algorithmic agents. Strong arguments support the position that algorithmic agents that operate without proper, or flawed, human oversight; or absent of well-defined governance and ethical frameworks, may have negative effects on greater societal norms and values such as the holy triumvirate of liberte, egalite, fraternite--or to put it in the language of the existing legal frameworks, fundamental human rights and freedoms, equality, and social cohesion. (14)

Responding to these background conditions, the University of Amsterdam launched a research program to study the effects of, and the normative considerations around, online personalized services in the domains of news, politics, commerce, and health communication. (15) This project has four major aims. First, we hope to identify how algorithmic agents tailor news, political communications, commercial offerings, and health-related information. Second, we want to understand what happens in the personalized and private information cocoons: what information individuals are exposed to and how they interact with algorithmic agents and their recommendations. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.