%0 Journal Article %1 hesmondhalgh2023impact %A Hesmondhalgh, David %A Campos Valverde, Raquel %A Kaye, D. Bondy Valdovinos %A Li, Zhongwei %D 2023 %I SSRN %J UK Centre for Data Ethics and Innovation Reports %K music systems recommendation algorithm from:kamber %P 58 %T The Impact of Algorithmically Driven Recommendation Systems on Music Consumption and Production: A Literature Review %X This report identifies two main bodies of academic research that pay sustained attention to algorithmic recommendation in the realm of culture: a) academic computer science and b) critical social science and humanities research on socio-technical systems, and discusses the strengths and weaknesses of each, pointing to the lack of collaboration between these two approaches. The report also shows there is very little publicly available research that examines in any depth the actual design and operation of music recommender systems (MRS), and their use by music streaming platforms (MSPs), as opposed to idealised models and experiments. There is very little sustained, publicly available research that examines the impact of MRS on music markets and experience, and how consumers/users of MSPs understand systems of recommendation. We also point to how research published by academic computer scientists and employees of tech companies on problems of MRS makes frequent use of the concepts of fairness and bias (most notably “popularity bias” and “demographic bias”, especially gender) but that critical social science and humanities research identifies potential limitations of these concepts, focusing instead on structural (in)justice and (in)equality. Problems of algorithmic opacity, transparency and accountability have been very widely discussed, often using the “black box” metaphor, and some researchers have identified simplifications and distortions in some uses of these terms. However, a more qualified and sophisticated notion of opacity may be appropriate for understanding the current way in which MRS operate.

@article{hesmondhalgh2023impact, abstract = {This report identifies two main bodies of academic research that pay sustained attention to algorithmic recommendation in the realm of culture: a) academic computer science and b) critical social science and humanities research on socio-technical systems, and discusses the strengths and weaknesses of each, pointing to the lack of collaboration between these two approaches. The report also shows there is very little publicly available research that examines in any depth the actual design and operation of music recommender systems (MRS), and their use by music streaming platforms (MSPs), as opposed to idealised models and experiments. There is very little sustained, publicly available research that examines the impact of MRS on music markets and experience, and how consumers/users of MSPs understand systems of recommendation. We also point to how research published by academic computer scientists and employees of tech companies on problems of MRS makes frequent use of the concepts of fairness and bias (most notably “popularity bias” and “demographic bias”, especially gender) but that critical social science and humanities research identifies potential limitations of these concepts, focusing instead on structural (in)justice and (in)equality. Problems of algorithmic opacity, transparency and accountability have been very widely discussed, often using the “black box” metaphor, and some researchers have identified simplifications and distortions in some uses of these terms. However, a more qualified and sophisticated notion of opacity may be appropriate for understanding the current way in which MRS operate.}, added-at = {2024-03-27T09:24:00.000+0100}, author = {Hesmondhalgh, David and Campos Valverde, Raquel and Kaye, D. Bondy Valdovinos and Li, Zhongwei}, biburl = {https://www.bibsonomy.org/bibtex/26d02bd25de846377b3ad54a71c85abc0/sop2-ffzg}, interhash = {2e42a7bc0c642b5d5217d8553b1673f9}, intrahash = {6d02bd25de846377b3ad54a71c85abc0}, journal = {UK Centre for Data Ethics and Innovation Reports}, keywords = {music systems recommendation algorithm from:kamber}, language = {English}, location = {https://ssrn.com/paper=4365916}, pages = 58, publisher = {SSRN}, timestamp = {2024-03-27T09:24:00.000+0100}, title = {The Impact of Algorithmically Driven Recommendation Systems on Music Consumption and Production: A Literature Review}, type = {Accepted Paper Series}, year = 2023 }