Revista ComHumanitas, ISSN: 1390-776X

Vol. 15, núm. 2 (2024), Julio - Diciembre 2024

DOI: https://doi.org/10.31207/rch.v15i2.456 

 

 

 

Personalización de contenidos en servicios de vídeo bajo demanda: Satisfacción con los algoritmos de las redes sociales

 

Personalization of Content in Video-on-Demand Services: Insights from Satisfaction over Social Media Algorithms

 

Personalização de conteúdo em serviços de vídeo sob demanda: satisfação com algoritmos de redes sociais

 

Alicia Urgellés-Molina[1]

Universidad Hemisferios

aliciau@uhemisferios.edu.ec

 

Mónica Herrero[2]

Universidad de Navarra

moherrero@unav.es

 

Fecha de recepción: 1 de diciembre de 2024

Fecha de aprobación: 12 de diciembre de 2024

Fecha de publicación: 17 de diciembre de 2024

Resumen

La rápida evolución de la personalización algorítmica ha reconfigurado la experiencia de los usuarios en las redes sociales y las plataformas de video a la carta, adaptando las recomendaciones de contenidos a las preferencias individuales. Este artículo explora la intersección entre la satisfacción del usuario, la capacidad de respuesta algorítmica y los patrones de consumo, haciendo hincapié en la influencia de las recomendaciones basadas en IA. Destaca cómo la percepción de la personalización algorítmica influye en el comportamiento del usuario, fomentando tanto el compromiso como la preocupación por las cámaras de eco y la privacidad.

Al examinar el paso de la emisión lineal tradicional al consumo no lineal de video a la carta, el estudio analiza cómo afectan los sistemas de recomendación a la calidad de la experiencia (QoE, por sus siglas en inglés). Además, aborda fenómenos como el binge-watching, la conciencia algorítmica y las implicaciones éticas de la personalización en la entrega de contenidos. Las diferencias generacionales en el uso de los medios ponen de relieve la interacción dinámica entre los hábitos de las redes sociales y las expectativas del vídeo a la carta.

Los resultados subrayan la necesidad de transparencia, diseño ético de la IA y alfabetización algorítmica para equilibrar la satisfacción del usuario con las prácticas responsables de los medios de comunicación. Esta investigación contribuye a comprender las implicaciones más amplias de la personalización en los ecosistemas de contenidos digitales, ofreciendo orientación para futuros desarrollos en los sistemas algorítmicos de medios de comunicación.

Palabras clave: Personalización algorítmica, vídeo a la carta, satisfacción del usuario, sistemas de recomendación, calidad de la experiencia

 

Abstract

The rapid evolution of algorithmic personalization has reshaped user experiences on social media and video-on-demand (VOD) platforms, tailoring content recommendations to individual preferences. This paper explores the intersection of user satisfaction, algorithmic responsiveness, and consumption patterns, emphasizing the influence of AI-driven recommendations. It highlights how perceptions of algorithmic customization impact user behavior, fostering both engagement and concerns over echo chambers and privacy.

By examining the shift from traditional linear broadcasting to non-linear VOD consumption, the study analyzes how recommendation systems affect Quality of Experience (QoE). Additionally, it addresses phenomena like binge-watching, algorithmic awareness, and the ethical implications of personalization in content delivery. Insights from generational differences in media use further underscore the dynamic interplay between social media habits and VOD expectations.

The findings underscore the need for transparency, ethical AI design, and algorithmic literacy to balance user satisfaction with responsible media practices. This research contributes to understanding the broader implications of personalization on digital content ecosystems, offering guidance for future developments in algorithmic media systems.

Keywords: Algorithmic Personalization, Video-on-Demand, User Satisfaction, Recommendation Systems, Quality of Experience (QoE)

 

Resumo

A rápida evolução da personalização algorítmica reformulou as experiências dos usuários nas plataformas de mídia social e de vídeo sob demanda (VOD), adaptando as recomendações de conteúdo às preferências individuais. Este artigo explora a interseção da satisfação do usuário, a capacidade de resposta algorítmica e os padrões de consumo, enfatizando a influência das recomendações orientadas por IA. Ele destaca como as percepções da personalização algorítmica afetam o comportamento do usuário, promovendo tanto o envolvimento quanto as preocupações com câmaras de eco e privacidade.

Ao examinar a mudança da transmissão linear tradicional para o consumo não linear de VOD, o estudo analisa como os sistemas de recomendação afetam a qualidade da experiência (QoE). Além disso, ele aborda fenômenos como binge-watching, consciência algorítmica e as implicações éticas da personalização no fornecimento de conteúdo. As percepções das diferenças geracionais no uso da mídia destacam ainda mais a interação dinâmica entre os hábitos de mídia social e as expectativas de VOD.

As descobertas ressaltam a necessidade de transparência, design ético de IA e alfabetização algorítmica para equilibrar a satisfação do usuário com práticas de mídia responsáveis. Esta pesquisa contribui para a compreensão das implicações mais amplas da personalização nos ecossistemas de conteúdo digital, oferecendo orientação para futuros desenvolvimentos em sistemas de mídia algorítmica.

Palavras-chave: Personalização algorítmica, vídeo sob demanda, satisfação do usuário, sistemas de recomendação, qualidade da experiência (QoE)

 

 

 

Introduction

The rise of algorithm-driven personalization has transformed how content is curated and consumed in digital spaces, especially on social media and video-on-demand (VOD) platforms. Algorithms analyze user behavior—such as likes, comments, and viewing habits—to tailor content according to perceived preferences. While many studies focus on the influence of these algorithms, findings are often inconsistent due to reliance on observational methods rather than controlled experiments. The few random experiments available, primarily conducted by companies that develop these algorithms, are challenging to replicate, limiting external researchers' ability to fully understand the mechanisms at play. There is a clear need for more randomized, independent studies to evaluate how these algorithms affect user behavior, particularly from researchers outside the companies that design these systems.

A deeper understanding of user perceptions of these algorithms is crucial, as it shapes how users engage with platforms. Users often assume that algorithms work based on personal engagement metrics—content they have liked, commented on, or watched previously. This belief is part of a broader feedback loop where user interactions continuously reshape the algorithm’s recommendations. This dynamic relationship means that users adjust their behavior based on what they perceive the algorithm’s preferences to be, feeding new data back into the system.

Having in mind the evolution of audiovisual content providers and audiovisual consumption, as it was briefly explained earlier, this research seeks to delve into the influence that consumption patterns on social media, strongly shaped by the filter of algorithmic recommendation, can have on audiences’ behavior in video-on-demand platforms. As algorithms curate and manage most content distribution on social media, it becomes important to understand the relationship that develops between users and their content filters. These systems have come to develop certain expectations and behaviors within media users and are molding the paths to media consumption over all other platforms.

With a special focus on the generational differences regarding social media consumption patterns and experiences, this exploratory paper provides a theoretical analysis of how Artificial Intelligence is transforming content personalization and its role within the Quality of Experience (QoE), as well as its perception by VOD users. We examine existing literature on AI-based recommendation algorithms, the factors that constitute QoE in VOD platforms, as well as insights from social media algorithms, highlighting their perceived functionality, and effectiveness in tailoring content to user preferences.

The paper also discusses the theoretical implications of these algorithms on user experience, including alleged increased satisfaction and platform engagement. In it, we will attempt to deepen into both the risks and opportunities personalization through AI might bring.  One of the risks is related, for example, to the fears associated with the use of personal data and the perception of surveillance surrounding AI and machine learning, since there seems to be no way for users to turn off the filtering (Van Den Bulck & Moe, 2017). Other risks come from the possibility of manipulations that could be conducted through them, since algorithms are not open to scrutiny (Striphas, 2015) and most organizations keep them enclosed. However, some authors think algorithms reduce cognitive bias. In fact, there are some studies that show that people tend to rely on the advice proposed by algorithms more than on human advice (Gil De Zúñiga et al., 2022).

Furthermore, we will also point out one of the risks concerning Quality of Experience which may not commonly be associated with it, as is pluralism and the concern for polarization. The non-existence of the raw algorithm, and the rationality behind its configuration means that it is possible to intervene in the design of the algorithm in a curatorial way, as it is expected in cultural industries. Morris (2015) takes Bourdieu’s concept of cultural intermediaries and considers recommendation systems as a new type of cultural intermediary that prepares cultural products for circulation with the help of algorithms: the way in which algorithmic recommendations frame content and manage their presentation would become an important part of the intermediation process of presentation and representation of culture and therefore, the production of culture. Therefore, personalization could also imply broadening the interests of audiences rather than simply adjusting the recommendations to their previous preferences, and that means a greater quality of video consumption experience. 

In any case, it seems audiences’ awareness towards algorithms may be crucial. As Zarouali (2021) points out, “the lack of algorithmic awareness might contribute to major societal problems, such as the spread of mis- and disinformation, the proliferation of filter bubbles, an increased susceptibility to data-driven manipulation, and the reinforcement of stereotypes, inequalities and discrimination”. 

Finally, this paper addresses the managerial challenges associated with data collection and usage, drawing from a wide range of academic sources to explore how these issues have been conceptualized and addressed across both social media and video-on-demand contexts. This research aims to synthesize existing knowledge and identify gaps for future exploration in the field of AI-driven personalization in media services.

Some key concepts for this research revolve around algorithm awareness, its relation to content and platform enjoyment, and the need for orientation these systems might fulfill for users.

Evolution of the Audiovisual Sector

The last decade has seen a significant evolution in the audiovisual sector, driven by the exponential growth of online content. The abundance of video content available has dramatically increased, offering users an unprecedented variety of viewing choices. Despite this proliferation, the amount of time that audiences can dedicate to viewing—arguably the most valuable resource in the media industry—remains limited.

One of the earliest solutions to navigating this content abundance was the Electronic Program Guide (EPG). Emerging in the United States as a cable television feature, the EPG provided updates on TV channels and schedules, helping users find content of interest. This system sought to solve the challenge of content discovery in a landscape with numerous channels, each with different schedules. The essence of the problem was information overload—finding the right program at the right time amidst a multitude of options.

As the media landscape shifted from multichannel TV providers to online VOD platforms, new challenges and opportunities emerged. The move to online platforms has fundamentally redefined consumption times and spaces, enabling a transition from a linear, mass-consumption model to a personalized experience more akin to browsing a digital library or shopping online. Users are no longer constrained by rigid schedules; they can access content on-demand, anytime and anywhere, according to their preferences.

This shift toward non-linear, individualized consumption—free from traditional constraints of time and space—aligns more closely with how other market goods are consumed. Unlike the linearity and simultaneous consumption associated with traditional broadcasting, online VOD consumption offers flexibility. Traditional TV was characterized by immediacy and the fleeting nature of live broadcasts, where content could not easily be revisited. Today, VOD platforms empower users to rewatch, pause, or skip through content at their convenience.

Belk (1975) discusses the importance of considering the circumstances, contexts, and situations surrounding consumers when studying their behavior and decision-making. A consumption context refers to the circumstances in which a consumer purchases or uses a product or service. Translating the concept of a catalog to online video on demand leads us to consider issues related to the redefinition of consumption times and spaces, as we move from a grid designed for linear and mass consumption to a situation of personalized consumption, more akin to that of a library or online commerce.

Users may perceive added value in their relationship with the television network simply from this transformation, but managing the change will be essential to leverage and offer a redefinition of the traditional service. We aim to highlight what can be incorporated into the traditional value proposition of the television network as it transitions to an environment that functions more like online retailing, a space where algorithmic recommendation and the application of AI are also of significant interest.

User Perceptions of Algorithmic Responsiveness and the Role of Recommendation Systems

Users often believe that social media and VOD algorithms understand their personal identities, adjusting content recommendations accordingly. These perceptions influence how users interact with platforms, expecting the algorithms to offer content that aligns with their interests and self-concept. For example, users might think that the content they see is a reflection of their past behaviors and preferences, such as videos they have watched or liked.

This belief system leads users to adjust their online behaviors to align with what they think the algorithm "wants" to see, essentially feeding the algorithm with behavior that reinforces their identity. As users perceive algorithms as more responsive to their needs and identities, their satisfaction with the platform increases. This is because users feel that the content provided is tailored specifically to them. However, this personalization can also have negative effects, such as reinforcing narrow viewpoints and limiting exposure to diverse perspectives.

Algorithmic responsiveness plays a significant role in shaping the user experience, with some platforms perceived as being more attuned to individual needs. TikTok, for example, is often viewed as more sensitive to user identity due to its algorithmic design, which focuses heavily on personalized content over social connections. This can lead to higher satisfaction among users who feel that their interests are better represented on such platforms. Understanding these perceptions is key to analyzing how user-algorithm interactions evolve over time and influence media consumption patterns.

Regarding consumption of experience goods, consumers usually turn to various sources for quality information on the product. Recommendation systems and engines are one of the ways in which media users can obtain information about the products they have available for consumption. They serve to reduce the consumer search costs and uncertainties associated with choosing unfamiliar products, thus facilitating online decision-making and engaging audiences more effectively. These algorithmically driven systems are central to search engines, social media platforms, and diverse content aggregators such as Amazon, YouTube, Spotify, and Netflix, not just video content.

The power of recommendation algorithms extends beyond VOD platforms to social media, where they similarly curate user experiences by presenting content most likely to keep users engaged. This process is central to the success of platforms like YouTube and TikTok, where users rely on recommendations to discover new videos, music, and trends. As algorithms analyze user behavior to identify patterns and preferences, they shape the pathways through which content is consumed, often introducing users to content they might not have actively sought out.

While on-demand consumption through OTT platforms has already somewhat disrupted the linearity of traditional consumption and generated certain behavioral changes in the audience, social networks like TikTok might be once again altering the relationship with content in terms of transience. As Muñoz-Gallego et al. (2024) explain, the conditions of brevity and non-permanence are accentuated on this social media network, which, thanks to its algorithm, is considered perhaps "the most dynamic platform today, with its low attention demand being comparable only to the old practice of channel surfing" (p.4).

Another way in which the consumption experience (and its perception) has been altered in audiovisual consumption is through the phenomenon of "binge-viewing": the act of watching several consecutive episodes of a series or films in a "binge". This form of consumption blurs the traditional rules of serialized productions and is another practice where the experience on social media can become quite similar to that of VOD platforms. The term "binge" has always had a negative connotation, associated with eating disorders or drinking, so the rapid normalization of the term for video consumption on the Internet might seem ironic. This option is presented to and perceived by users as a form of relaxation and release (Feeney, 2014; Skipper, 2014), similar, we could point out, to the hours spent continuously swiping up on social media.

Algorithmic recommendation leverages AI algorithms to analyze large datasets on user behavior, such as viewing history, preferences, and interactions, to create personalized experiences. QoE and algorithmic recommendation are closely connected, as content personalization through algorithms can significantly enhance QoE (Taylor and Choi, 2022). Quality of Experience (QoE) is a multifaceted concept that encompasses various aspects of a user's interaction with a service or product. AI-driven content personalization plays a crucial role in enhancing QoE by reducing information overload (Zarouali et al., 2021).

AI can filter out irrelevant content and present users only with what interests them, minimizing information overload and frustration. AI can also help by improving discoverability; AI can help users find new content they might not have encountered otherwise, broadening their horizons and providing a richer experience. A study (Hong, 2023) found that satisfaction with Netflix's algorithmic recommendations was comparable to that of personal recommendations. This creates an overall more intuitive experience, with AI algorithms learning from user interactions to deliver a more user-friendly experience (Jayanthiladevi et al., 2020).

Algorithmic recommendation can impact QoE by shaping the content users see in their feeds. The perception of algorithmic responsiveness (PAR)—the belief that the algorithm understands and validates the user’s identity—has been linked to greater enjoyment of social media. In a study, TikTok scored higher in PAR than Facebook, suggesting that TikTok's algorithm was perceived as more attuned to the user’s identity (Taylor and Choi, 2022). Another study looked at how Twitter's recommendation algorithms affect users' emotions, such as anger, sadness, anxiety, and happiness. They found that Twitter's engagement algorithm increased exposure to content that expressed negative emotions (Milli et al., 2023).

It is important to note that QoE is a subjective concept and can vary significantly from one user to another. In addition to the aspects mentioned above, there are other factors that can influence QoE, such as the technical quality of the service (e.g. transmission speed, video resolution) and the context of use (e.g. location, device).

Risks and Opportunities of AI-Driven Personalization

The integration of AI in content personalization offers significant opportunities, such as improved user satisfaction and engagement. However, it also comes with inherent risks. Concerns about privacy, data security, and algorithmic transparency have grown as platforms increasingly use personal data to refine recommendations. Users often feel uneasy about the lack of control over their data and the opacity of algorithmic processes.

For instance, users may be unaware that there is no option to disable algorithmic filtering on most platforms. This lack of transparency fosters concerns about surveillance and data manipulation. Authors like Van Den Bulck and Moe (2017) highlight the potential for AI to perpetuate biases or reinforce echo chambers, as algorithms can subtly shape user perceptions without their knowledge. Striphas (2015) suggests that algorithms' role in content distribution raises ethical questions, given that most are proprietary and lack external oversight.

Conversely, some studies argue that algorithmic recommendations can counteract certain cognitive biases by providing diverse suggestions. Research by Gil De Zúñiga et al. (2022) shows that users may trust algorithmic advice more than human recommendations, especially when selecting content. This trust underscores the importance of examining how AI shapes user behavior on VOD platforms and the implications for media consumption.

Another behavior that algorithmic personalization has an impact on is binge viewing. In the context of television, a classic way of measuring the relationship between channel and audience is through the time spent in front of the screen (Danaher & Lawrie, 1998). The time spent watching content from a channel indicates, in principle, the satisfaction with that offer (Phalen & Ducey, 2012). Applying this perspective used in engagement studies, which is mainly aimed at the amount of time viewers spend watching content and the attention with which they do so, continuous viewing and binge viewing could be seen as a positive behavior. The term "binge", however, carries a historical association with excess and loss of control, often linked to eating or drinking disorders. Yet, in the context of media consumption, it has gained a more normalized meaning, representing a common form of relaxation and entertainment. This shift in perception highlights the role of VOD platforms in redefining how audiences interact with long-form content.

Binge-watching has become a defining feature of modern media consumption, challenging the traditional rules of serialized production. This behavior reflects the evolving user expectations in a VOD environment where content is accessible on-demand. Unlike linear TV, where audiences had to wait for weekly episodes, VOD platforms offer entire seasons at once, allowing users to consume content at their preferred pace.

Studies have shown that users perceive binge-watching as a form of escapism  or a way to unwind, similar to the experience of scrolling through social media feeds for extended periods. The ability to control one's viewing experience—pausing, resuming, or skipping episodes—contributes to a sense of empowerment and personalization. However, this form of consumption also raises questions about its impact on mental well-being and how platforms balance user satisfaction with responsible content delivery.

Challenges in Managing Data and User Perceptions: The Role of Algorithmic Awareness

Managing user data is a complex task for VOD platforms, which must balance personalization with privacy concerns. Transparent data practices are essential for building user trust, as is providing users with control over their information. However, achieving this balance can be challenging, as platforms must navigate the tension between offering highly tailored experiences and respecting user privacy. These challenges encompass technical, financial, ethical, and regulatory aspects.

A crucial part of this challenge involves addressing the risks of "filter bubbles" and polarization. Algorithms that prioritize user preferences can inadvertently limit exposure to diverse content, creating an environment where users are repeatedly shown similar types of content. This can reinforce existing biases and reduce opportunities for discovering new perspectives. Understanding how users perceive the balance between personalization and diversity is key to improving QoE on VOD platforms.

Rostamiani and Moradi Kamreh (2024) point out how the integration of AI technologies into existing media infrastructures is technically challenging. It often requires significant financial investment to implement and maintain AI systems while ensuring their reliability and scalability.

In addition to these challenges, media organizations must also consider the need for algorithmic literacy: A deeper understanding of how algorithms function in the news selection process and their non-neutral nature will enable a more critical consumption of algorithm-driven news (Gil De Zúñiga et al., 2022).

Algorithmic awareness —users' understanding of how algorithms shape their content— plays a crucial role in shaping their interactions with platforms. Studies suggest that most users have a limited grasp of how recommendation systems operate, often misunderstanding the mechanisms behind content selection. This gap in understanding can lead to confusion and dissatisfaction when users encounter content that seems misaligned with their interests.

Zarouali (2021) highlights the societal risks of low algorithmic awareness, including the spread of misinformation, filter bubbles, and susceptibility to manipulation. Enhancing algorithmic literacy through education could empower users to make more informed decisions about their media consumption. Such efforts could mitigate some of the negative effects of AI-driven personalization, fostering a healthier relationship between users and recommendation systems.

 

Conclusion and future directions

The ongoing evolution of VOD platforms and their integration with AI-based recommendation algorithms presents both opportunities and challenges. Understanding the dynamics of user-algorithm interactions is essential for designing systems that enhance user satisfaction while promoting transparency and diversity. As social media and VOD platforms continue to intersect, examining the influence of social media consumption on VOD behavior becomes increasingly relevant.

Future research should focus on developing robust metrics for user satisfaction, exploring ethical frameworks for algorithm design, and investigating the long-term effects of personalized recommendations on media consumption patterns. Additionally, studies should address generational differences in algorithmic perceptions, as younger users might have different expectations for personalization compared to older audiences.

By synthesizing insights from the literature on social media algorithms and VOD platforms, this study aims to contribute to a deeper understanding of how AI is reshaping the landscape of digital content consumption. Addressing these questions will be crucial for the future of media services, ensuring that AI-driven personalization is used responsibly and ethically to enhance the user experience.

 

References

Baraković Husić, J., Baraković, S., Cero, E., Slamnik, N., Oćuz, M., Dedović, A., & Zupčić, O. (2020). Quality of experience for unified communications: A survey. International Journal of Network Management, 30(3), 1–25. https://doi.org/10.1002/nem.2083

Belk, R. W. (1975). Situational Variables and Behavior Consumer. Journal of Consumer Research, 2(3), 157-164.

Danaher, P. J., & Lawrie, J. M. (1998). Behavioral Measures of Television Audience Appreciation. Journal of Advertising Research, 1, 54-65.

Dogruel, L., Facciorusso, D., & Stark, B. (2022). ‘I’m still the master of the machine.’Internet users’ awareness of algorithmic decision-making and their perception of its effect on their autonomy. Information, Communication & Society, 25(9), 1311–1332.

Feeney, N. (2014). When, exactly, does watching a lot of Netflix become a ‘binge.’ The Atlantic, 18.

Gil De Zúñiga, H., Cheng, Z., & González-González, P. (2022). Effects of the news finds me perception on algorithmic news attitudes and social media political homophily. Journal of Communication, 72(5), 578–591. https://doi.org/10.1093/joc/jqac025

Hong, S. (2023, April 21). Believe in algorithms or believe in your friends?: the influences of the different ways in seeking what to watch on Netflix. https://repositories.lib.utexas.edu/items/ba91cb1a-8fa5-458f-ad5d-4b4c7e0e12d0

Jayanthiladevi, A., Raj, A. G., Narmadha, R., Chandran, S., Shaju, S., & Krishna Prasad, K. (2020). AI in Video Analysis, Production and Streaming Delivery. Journal of Physics: Conference Series, 1712(1). https://doi.org/10.1088/1742-6596/1712/1/012014

Milli, S., Carroll, M., Wang, Y., Pandey, S., Zhao, S., & Dragan, A. D. (2023). Engagement, User Satisfaction, and the Amplification of Divisive Content on Social Media. http://arxiv.org/abs/2305.16941

Morris, J. W. (2015). Curation by code: Infomediaries and the data mining of taste. European Journal of Cultural Studies, 18(45), 446–463.

Muñoz-Gallego, A., Giri, L., Nahabedian, J. J., & Rodríguez, M. (2024). Audiovisual Narratives on Tik Tok: New Challenges for Public Communication of Science and Technology. Revista Mediterranea de Comunicacion, 15(1), 145–161. https://doi.org/10.14198/MEDCOM.25481

O’Sullivan, Derry et al. (2024). Improving the Quality of the Personalized Electronic Program Guide. User modeling and user-adapted interaction 14(1), 5–36.

Phalen, P. F., & Ducey, R. V. (2012). Audience Behavior in the Multi-Screen «Video-Verse». International Journal on Media Management, 14(2), 141-156. https://doi.org/10.1080/14241277.2012.657811

Rostamian, S., & Moradi Kamreh, M. (2024). AI in Broadcast Media Management: Opportunities and Challenges. AI and Tech in Behavioral and Social Sciences, 2(3), 21-28. https://doi.org/10.61838/kman.aitech.2.3.3

Skipper, B. (2014). House of cards: Will Netflix’s binge-viewing approach to TV become the norm. International Business Times, 14.

Striphas, T. (2015). Algorithmic culture. European Journal of Cultural Studies, 18 (45), 395–412.

Taylor, S. H., & Choi, M. (2022). An Initial Conceptualization of Algorithm Responsiveness: Comparing Perceptions of Algorithms Across Social Media Platforms. Social Media and Society, 8(4). https://doi.org/10.1177/20563051221144322

Van Den Bulck, H., & Moe, H. (2017). Public service media, universality and personalisation through algorithms: mapping strategies and exploring dilemmas. Media, Culture and Society, 40 (60), 1–18.

Zarouali, B., Boerman, S. C., & de Vreese, C. H. (2021). Is this recommended by an algorithm? The development and validation of the algorithmic media content awareness scale (AMCA-scale). Telematics and Informatics, 62. https://doi.org/10.1016/j.tele.2021.101607