Year: 2024
Type: dissertation
Abstract: Digital services use recommender algorithms to sort and recommend content based on user preferences. Many public service media companies have adapted the use of recommender algorithms in digital news services. Algorithmic data gathering involves ethical aspects concerning media companies and service users. This research investigates how young users understand news recommender algorithms by focusing on the case of Finland’s public service media company Yleisradio's (Yle) news application Uutisvahti's recommender system Voitto robot. The study explores algorithmic understanding by focusing on the themes of algorithmic awareness, algorithmic transparency and user trust in recommender systems. The research was implemented among 16 to 18-year-old expatriate Finns in a Finnish upper secondary school in Spain. Qualitative research was implemented using mixed methods and thematic analysis. User data was gathered with a background questionnaire and face-to-face pair and group interviews. Fifteen students participated in the questionnaire and thirteen in the interviews. The main research findings indicate three levels of algorithmic understanding: users who do not understand algorithms, users whose understanding increased during the study and users who understand algorithms. The most significant factor affecting users' perception of their algorithmic understanding was the impact of the feedback loop between the user and the algorithm. Negative feedback loops increased algorithmic awareness and critical thinking in users but did not increase users' algorithmic understanding and user trust. Users had concerns of the ethical aspects of algorithms and algorithmic data gathering. The research results indicate that algorithmic transparency supports youth understanding and user trust in data gathering and the recommender system. Algorithmic understanding can be supported by providing information, education, diverse experiences and possibilities for discussion and reflection. Youth algorithmic understanding can be increased by resolving the issues causing negative feedback loops, providing more user autonomy and developing the interaction of the recommendation system.
Author: Kauramäki, Laura
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7132977934