The Brokenness in Our Recommendation Systems: Computational Art for an Ethical Use of A.I.

Abstract

AbstractOnline recommendation systems are information filtering systems that provide users with streams of prioritized content based on expected individual preferences. While they can be of different types—e.g. collaborative, content-based, or hybrid filtering, they typically share the use of machine learning as a type of artificial intelligence able to perform predictions and profile personal taste. Drawing upon previous research in critical algorithm studies, this paper tackles the limitations of predictive content personalization and automated sorting. It does so by taking as a case study the computational art project This Recommendation System is Broken, developed in collaboration with metaLAB (at) Harvard as part of Curatorial A(i)gents (2020–2022), a series of experiments at the Harvard Art Museums exploring the interplay between A.I. and curatorial practices. While advocating for the ethical design and use of artificial intelligence, I discuss creative coding (Maeda in Creative code, Thames & Hudson Inc., New York, 2004 [1]) as a mode of engagement for artists, designers, media practitioners to take action in the development of context-sensitive algorithms that promote speculative (Dunne and Raby in Speculative everything: design, fiction and social dreaming. The MIT Press, 2013 [2]) design practices and sustainable future-making (Yelavich and Adams in Design as future-making. Bloomsbury Academic, London, GB, 2014 [3]), rather than adopting predictive or statistical models. The art project presented here challenges us to consider the biases of automated decision-making in generating instances of visibility/invisibility on media platforms and other online environments. What we might call “brokenness” is ultimately an attempt to escape the illusory quest for certainty and artificial perfection. Even more so, it is about shaping an ethic of A.I. practice and understanding how information filtering systems are transforming contemporary media cultures.KeywordsComputational artAlgorithmsA.I. ethicsOnline platformsRecommendation systems