I am currently carrying out a 4-year research project “Do algorithms know better? First-person authority in the age of big data”. The project is funded by the Netherlands Organisation for Scientific Research (NWO, Veni).
The project departs from an assumption firmly rooted in modern liberal values and practices, which is that individuals are authoritative about their own attitudes. In recent years, however, this principle of ‘first-person authority’ has come under pressure by what I call ‘algorithmic authority’, or the growing knowledge and predictive power algorithms, and the institutions making use of them, have about us. This development leads to inevitable tensions between what individuals think they want, value and intend, and what algorithms predict they want, value and intend. These tensions are reinforced by the fact that algorithms have a potentially transformative effect in shaping – or manipulating - people’s attitudes, goals, and values.
The project consists of: (1) a conceptual stage, in which address the epistemic question in what sense individuals and algorithms can both be authoritative or ‘know best’, (2) an evaluative stage in which I analyse what is at stake, morally, when algorithmic predictions are taken as proxy for individuals’ attitudes and in what sense online manipulation would undermine autonomy, and (3) a policy-attuned stage (during which I will collaborate with the Rathenau Institute), where I address the question in which specific cases and under what circumstances deferring to algorithmic authority would be justified, and how the potential benefits of doing so must to be weighed against the need to respect people’s first-person authority and autonomy.