Shannon Brick

Algorithms & Authenticity
[Under review]

Algorithmic Decision Aids (ADAs) are tools that help people make decisions more accurately, more efficiently, or both. This paper distinguishes between two versions of the claim that ADAs threaten personal authenticity. The first version of the worry is that ADAs threaten authenticity because they threaten our ability to put our values into action. An agent committed to fairness, for instance, risks deciding in a way that is unfair when she recruits an ADA, because the ADA is liable to be biased in ways she cannot easily see. On the second version of the worry, ADAs threaten authenticity because they risk undermining the ability of our decisions to express our understanding of our reasons for action. This is a problem because this variety of authenticity is essential for certain relational goods, like friendship, intimacy and (in some cases) a healthy democracy. While both versions of the worry are serious, the paper argues that the first version of the worry is significantly more tractable than the second.

Biometric Attention-Tracking & Education (with Alicia Patterson)
[Revise & resubmit]

Recent advances in Artificial Intelligence have led to a new wave of optimism about the efficiency-enhancing prospects of educational technology. This paper critically analyses the promise of a nascent form of educational AI, dubbed Biometric Attention-Tracking (BAT) technology. BAT seeks to harness advances in Computer Vision to augment (or replace) a teacher’s capacity to monitor and evaluate student engagement in the classroom. The paper develops three independent lines of criticism. The first line challenges some of the central assumptions about student attention that underpin the current functioning of these systems, so as to demonstrate that these technologies are unlikely to deliver on what is promised of them (at least not any time soon.) The second and third lines of criticism start by imagining that these systems could deliver on their promise of measuring student attention accurately and reliably for all students. It is argued that, even if this were true, we would have strong pro tanto reasons to avoid deploying BAT, because doing so would pose serious risks to several educational goods. 

Digital Manipulation & Domination
[In preparation]

The first part of this paper motivates the claim that, when responding to concerns about digital manipulation, we should avoid centering the value of personal autonomy. The second part considers whether republicanism can provide a suitable alternative for illuminating, and substantiating, these concerns. It’s argued that to the extent that they focus on targeted advertising, existing attempts to use republicanism to answer normative questions about digital manipulation face serious limits. Nonetheless, republicanism might still be able illuminate problems with digital manipulation, so long as one focuses on features of the data economy that extend beyond targeted ads.