A Right to Constructive Optimization: A Public Interest Approach to Recommender Systems in the Digital Services Act

What the paper says

The technological promise of recommender systems should not be misused by those with decisional power over the infrastructural, data, and knowledge resources needed for their design. The ideal of personalization should not mask self-serving optimization. Instead, we propose that people, not only in their capacity as consumers but, more generally, as democratic citizens, have a legitimate claim to ensure that very large online platforms (or VLOPs) respect their interests within optimization processes through the content policy strategies and recommendation technologies they employ. To this end, this paper argues for, and develops, a right to constructive optimization that promotes people's effective enjoyment of fundamental rights and civic values in digital settings. The argument is structured as follows. First, the paper strengthens the claim that the largest online platforms perform a public function (although this is not the only way such functions can be performed). Second, drawing from the philosophy of Iris Marion Young, the paper identifies self-determination and self-development as key values recommenders should promote as part of this crucial function under conditions of inclusivity, political equality, reasonableness, and publicity. After having critiqued the EU Digital Services Act's approach toward regulating the function recommenders hold, the right to constructive optimization is concretized as an alternative normative benchmark and used as an interpretative lens to enrich ongoing legal initiatives.