YouTube Recommendations Research Articles

Auditing Entertainment Traps on YouTube: How Do Recommendation Algorithms Pull Users Away from News

ABSTRACT Recommendation algorithms that customize information feeds for individuals have raised concerns about exacerbating inequalities in news exposure among citizens. In response to these concerns, we conducted an audit study on YouTube to analyze the algorithmic impact on curating news versus other content topics. We examined over 1.7 million YouTube video recommendations audited in 2019 and developed novel analysis approaches including network analysis and Markov chains. Results show that recommendation algorithms may potentially redirect users away from news content through two influence pathways: (1) the “topical filter bubbles,” wherein entertainment content has a higher probability of being recommended over news content in a self-reinforcing manner; and (2) “algorithmic redirection,” wherein the probability of entertainment videos being recommended after a news video is much higher than that for the opposite. Overall, YouTube recommendation algorithms have a higher probability of recommending entertainment videos than news. The findings imply essential biases in algorithmic recommendations on digital platforms beyond amplifying users’ preferences.

Artificial Intelligence Applications Used in On-line Retail in China and Their Relationship to Customer Satisfaction and Loyalty

Artificial Intelligence is the creation of intelligent computers and smart computer algorithms that help machines understand human intelligence (IBM Cloud Education, 2021). Artificial Intelligence is created by analyzing behavior and patterns of big data. Artificial intelligence has existed since the 1950s (Song et al, 2019) but during recent times application of AI in various sectors like healthcare, business, entertainment, education, weather, and geology has picked up momentum. To cite a few well-known examples of Artificial Intelligence technology that are used by almost everyone in our day-to-day activities are advanced Google searches, YouTube recommendations, and Alexa (Song et al, 2019). The advent of technology has brought about innovation and transformation to every aspect of the business world, including the retail industry. The shift in the retail sector becomes obvious when big brick-and-mortar retailers scale down their physical shops and gradually move part of their business online. Consumer attitudes toward online shopping can vary based on demographics, perceived risk, perceived ease of use, and perceived usefulness. Demographic characteristics can further be classified as sex, age, educational qualification, household income, and relationship status, to name a few. In this study, the aim is to identify artificial intelligence applications and their use in online retail by Chinese consumers. A model has been developed to test the relationships between artificial intelligence applications used for online shopping and customer satisfaction and loyalty.

The filter bubble generated by artificial intelligence algorithms and the network dynamics of collective polarization on YouTube: the case of South Korea

ABSTRACT This study examined the role of the ‘filter bubble,’ an algorithm-mediated YouTube video suggestion system, in political polarization and the presence of echo chamber patterns in public engagement. We examined the mechanism by which automated YouTube recommendations augment selective exposure to ideologically similar content and the network-based dynamics of collective polarization. We collected lists of videos recommended by conservative and progressive news media and the accompanying replies using YouTube application programming interfaces (APIs) embedded in YouTube Data Tools and Webometric Analyst 2.0. The study examined similarities in content of related videos and conversation networks between commenters, as well as the relationships between the videos, channels, and commenters. Almost half of the videos had content similar to the original videos offered by both conservative and progressive news media. The filter bubbles in progressive media were not as strong as those in conservative media. More than 30% of the videos recommended by progressive media were from legacy media, while this figure was 25% for conservative sources. This study has significant implications for reverse engineering studies because it collected data through APIs, which are neutral and can access YouTube servers without logging into individual IDs.

Auditing YouTube's recommendation system for ideologically congenial, extreme, and problematic recommendations.

Algorithms of social media platforms are often criticized for recommending ideologically congenial and radical content to their users. Despite these concerns, evidence on such filter bubbles and rabbit holes of radicalization is inconclusive. We conduct an audit of the platform using 100,000 sock puppets that allow us to systematically and at scale isolate the influence of the algorithm in recommendations. We test 1) whether recommended videos are congenial with regard to users' ideology, especially deeper in the watch trail and whether 2) recommendations deeper in the trail become progressively more extreme and come from problematic channels. We find that YouTube's algorithm recommends congenial content to its partisan users, although some moderate and cross-cutting exposure is possible and that congenial recommendations increase deeper in the trail for right-leaning users. We do not find meaningful increases in ideological extremity of recommendations deeper in the trail, yet we show that a growing proportion of recommendations comes from channels categorized as problematic (e.g., "IDW," "Alt-right," "Conspiracy," and "QAnon"), with this increase being most pronounced among the very-right users. Although the proportion of these problematic recommendations is low (max of 2.5%), they are still encountered by over 36.1% of users and up to 40% in the case of very-right users.

Recommending instead of taking down: YouTube hyperpartisan content promotion amid the Brazilian general elections

AbstractYouTube released a global commitment to reduce the spread of problematic content by actively recommending “trusted” news sources on their platform but did not disclose the criteria used to classify the reputation of channels and the list of trusted news sources favored by the company. We aim to understand the patterns of recommendation by YouTube's algorithm and identify what the platform classifies as “reliable informative sources.” In our experiment, we performed 205 tests on YouTube's recommendation system, simulating the behavior of new Brazilian users with no historical data on the platform looking for reliable news throughout the elections. We documented the Brazilian media groups that were suggested on the homepage, further analyzing the presence, position and content of Jovem Pan videos, Brazil's flagship conservative media group. We identified that, far from reigning in extremist content, YouTube systematically prioritized Jovem Pan and did not actively take down toxic Jovem Pan content. Our findings indicate that YouTube magnified the imbalance between candidates' portrayals, exposing the regulatory asymmetry between broadcast media and online platforms in the country. In spite of the opaque commercial agreements and business strategies, the platform's recommendation decisions may have an “advertising effect” of sponsorship and endorsement on users.

A Utility-Preserving Obfuscation Approach for YouTube Recommendations

Online content platforms optimize engagement by providing personalized recommendations to their users. These recommendation systems track and profile users to predict relevant content a user is likely interested in. While the personalized recommendations provide utility to users, the tracking and profiling that enables them poses a privacy issue because the platform might infer potentially sensitive user interests. There is increasing interest in building privacy-enhancing obfuscation approaches that do not rely on cooperation from online content platforms. However, existing obfuscation approaches primarily focus on enhancing privacy but at the same time they degrade the utility because obfuscation introduces unrelated recommendations. We design and implement DeHarpo, an obfuscation approach for YouTube's recommendation system that not only obfuscates a user's video watch history to protect privacy but then also denoises the video recommendations by YouTube to preserve their utility. In contrast to prior obfuscation approaches, DeHarpo adds a denoiser that makes use of a ``secret'' input (i.e., a user's actual watch history) as well as information that is also available to the adversarial recommendation system (i.e., obfuscated watch history and corresponding ``nois`` recommendations). Our large-scale evaluation of DeHarpo shows that it outperforms the state-of-the-art by a factor of 2x in terms of preserving utility for the same level of privacy, while maintaining stealthiness and robustness to de-obfuscation.

Exploring YouTube's Recommendation System in the Context of COVID-19 Vaccines: Computational and Comparative Analysis of Video Trajectories.

Throughout the COVID-19 pandemic, there has been a concern that social media may contribute to vaccine hesitancy due to the wide availability of antivaccine content on social media platforms. YouTube has stated its commitment to removing content that contains misinformation on vaccination. Nevertheless, such claims are difficult to audit. There is a need for more empirical research to evaluate the actual prevalence of antivaccine sentiment on the internet. This study examines recommendations made by YouTube's algorithms in order to investigate whether the platform may facilitate the spread of antivaccine sentiment on the internet. We assess the prevalence of antivaccine sentiment in recommended videos and evaluate how real-world users' experiences are different from the personalized recommendations obtained by using synthetic data collection methods, which are often used to study YouTube's recommendation systems. We trace trajectories from a credible seed video posted by the World Health Organization to antivaccine videos, following only video links suggested by YouTube's recommendation system. First, we gamify the process by asking real-world participants to intentionally find an antivaccine video with as few clicks as possible. Having collected crowdsourced trajectory data from respondents from (1) the World Health Organization and United Nations system (nWHO/UN=33) and (2) Amazon Mechanical Turk (nAMT=80), we next compare the recommendations seen by these users to recommended videos that are obtained from (3) the YouTube application programming interface's RelatedToVideoID parameter (nRTV=40) and (4) from clean browsers without any identifying cookies (nCB=40), which serve as reference points. We develop machine learning methods to classify antivaccine content at scale, enabling us to automatically evaluate 27,074 video recommendations made by YouTube. We found no evidence that YouTube promotes antivaccine content; the average share of antivaccine videos remained well below 6% at all steps in users' recommendation trajectories. However, the watch histories of users significantly affect video recommendations, suggesting that data from the application programming interface or from a clean browser do not offer an accurate picture of the recommendations that real users are seeing. Real users saw slightly more provaccine content as they advanced through their recommendation trajectories, whereas synthetic users were drawn toward irrelevant recommendations as they advanced. Rather than antivaccine content, videos recommended by YouTube are likely to contain health-related content that is not specifically related to vaccination. These videos are usually longer and contain more popular content. Our findings suggest that the common perception that YouTube's recommendation system acts as a "rabbit hole" may be inaccurate and that YouTube may instead be following a "blockbuster" strategy that attempts to engage users by promoting other content that has been reliably successful across the platform.

YOUTUBE VIDEO RECOMMENDATION BASED ON FACIAL EXPRESSION

Human faces play a significant role in determining a person's mood. Emotion plays an important role in various fields, including biomedical engineering, brain science, neuroscience, psychological wellness, and mental health. Emotions are not only used for determination of the state of the human brain but too utilized as a proposal framework to help people in finding things that coordinate their needs and inclinations. This motivated us to create a framework that can viably and proficiently recognize feelings from the facial expressions of the user and recommend the YouTube videos based on that emotion. Emotions are bought by the neurophysiological changes that are associated with thoughts, feelings and behavioral changes in a person. Every emotion needs to be treated in a right way and research suggests that watching YouTube videos influences your mood, and in our experiment, we try to influence it in a good way. We can detect changes in emotion by observing facial expressions, body language, and tone of voice. There are many indicators of a person's mood, but facial expressions are one of the most important. In our experiment, we used the Haar Cascade Classifier to detect the face and a convolution neural network to analyze the facial expression. Performance of our proposed model is 95.66% for face detection and 61.88% for emotion detection. Key Words: Emotion Detection, FER-2013, HAAR Cascade, YouTube Recommendation, CNN.

YouTube's recommendation algorithm is left-leaning in the United States.

With over two billion monthly active users, YouTube currently shapes the landscape of online political video consumption, with 25% of adults in the United States regularly consuming political content via the platform. Considering that nearly three-quarters of the videos watched on YouTube are delivered via its recommendation algorithm, the propensity of this algorithm to create echo chambers and deliver extremist content has been an active area of research. However, it is unclear whether the algorithm may exhibit political leanings toward either the Left or Right. To fill this gap, we constructed archetypal users across six personas in the US political context, ranging from Far Left to Far Right. Utilizing these users, we performed a controlled experiment in which they consumed over eight months worth of videos and were recommended over 120,000 unique videos. We find that while the algorithm pulls users away from political extremes, this pull is asymmetric, with users being pulled away from Far Right content stronger than from Far Left. Furthermore, we show that the recommendations made by the algorithm skew left even when the user does not have a watch history. Our results raise questions on whether the recommendation algorithms of social media platforms in general, and YouTube, in particular, should exhibit political biases, and the wide-reaching societal and political implications that such biases could entail.

Exploring how a YouTube channel’s political stance is associated with early COVID-19 communication on YouTube

ABSTRACT The unexpected chaos of the COVID-19 pandemic brought about a host of political discussions in the US, including conversations about the Trump administration’s COVID-related policies. During these conversations, major news outlets represented each political group’s stance using social media in addition to their traditional channels. This study explores how those news outlets’ political stances relate to early COVID-19 social media discourse with three distinct methods. We selected two YouTube videos about the World Health Organization (WHO)’s declaration of a pandemic from two news outlets with contrasting political stances, Fox News and MSNBC. Three types of analyses were conducted to compare those videos on different observation levels: (1) a video network analysis based on the YouTube recommendation algorithm, (2) a framing analysis exploring how differently the two news outlets delivered the news, and (3) a content analysis of viewer comments. Our results demonstrate that the videos were not directly connected to each other in the YouTube recommendation network, that they delivered the same news with contrasting frames, and that characteristics of the viewers’ comments – commenters’ political positions, roles in knowledge sharing, and social cues – significantly or partially significantly related to the YouTube channel’s political position. Based on these findings, we argue that while news outlets’ political stances were associated with the behaviors of all three types of agents on YouTube – news outlets, their viewers, and the platform – to some extent, the viewer groups’ knowledge-sharing behaviors did not significantly differ as the news outlets’ contrasting frames did.

YouTube radical content recommendations: a meta-analysis

YouTube's recommendation system algorithms have been the subject of controversy and debate, as there have been incidents of the recommendation system offering hazardous, most often extremist content to users. The purpose of this study is to test the hypothesis that YouTube recommendations can contain and promote radical content. The scientific and practical significance of this work is that YouTube has a huge popularity, and knowledge about the algorithms of its recommendation system would help to improve information literacy of users and thus prevent the potential danger that can be caused by exposure to malicious information. This study seeks to popularize the idea that all information on the Internet requires a thorough verification. Improving information literacy of users, this research contributes globally to the prevention of terrorist attacks, acts of self-mutilation, suicide, pedophilia, etc. The sources for the present investigation with description at least one type of radical content were published in the last 5 years. The material of interest was searched through authoritative scientific and metric databases, namely: Google Scholar, Scopus, Web of Science, and PubMed by extracting suitable records. A total of 22 studies were pooled according to the eligibility and exclusion criteria, and a meta-analysis was performed which found that in 13 of them YouTube recommendations contained and promoted hazardous content, in 7 studies the scientists presented ambiguous results, and only in 2 studies the authors found that the recommendations did not contain or promote any radical content. Thus, according to the results of this research, it was found that YouTube recommendations can contain and disseminate prohibited hazardous content, and therefore the authors strongly recommend making adjustments to the algorithms of the recommendations in order to protect users from prohibited information, including the promotion of extremism and violence. Keywords: YouTube, recommendations, recommendation system, hazardous content, pseudoscientific content, radical content, extremism, pedophilia, meta-analysis.

The spread of COVID-19 vaccine information in Arabic on YouTube: A network exposure study.

The Arabic-speaking world had the lowest vaccine rates worldwide. The region's increasing reliance on social media as a source of COVID-19 information coupled with the increasing popularity of YouTube in the Middle East and North Africa region begs the question of what COVID-19 vaccine content is available in Arabic on YouTube. Given the platform's reputation for being a hotbed for vaccine-related misinformation in English, this study explored the COVID-19 vaccine-related content an individual is likely to be exposed to on YouTube when using keyword-based search or redirected to YouTube from another platform from an anti-vaccine seed video in Arabic. Only using the Arabic language, four networks of videos based on YouTube's recommendations were created in April 2021. Two search networks were created based on Arabic pro-vaccine and anti-vaccine keywords, and two seed networks were created from conspiracy theory and anti-vaccine expert seed videos. The network exposure model was used to examine the video contents and network structures. Results show that users had a low chance of being exposed to anti-vaccine content in Arabic compared to the results of a previous study of YouTube content in English. Of the four networks, only the anti-vaccine expert network had a significant likelihood of exposing the user to more anti-vaccine videos. Implications were discussed. YouTube deserves credit for its efforts to clean up and limit anti-vaccine content exposure in Arabic on its platform, but continuous evaluations of the algorithm functionality are warranted.

Radical bubbles on YouTube? Revisiting algorithmic extremism with personalised recommendations

Radicalisation via algorithmic recommendations on social media is an ongoing concern. Our prior study, Ledwich and Zaitsev (2020), investigated the flow of recommendations presented to anonymous control users with no prior watch history. This study extends our work on the behaviour of the YouTube recommendation algorithm by introducing personalised recommendations via personas: bots with content preferences and watch history. We have extended our prior dataset to include several thousand YouTube channels via a machine learning algorithm used to identify and classify channel data. Each persona was first shown content that corresponded with their preference. A set of YouTube content was then shown to each persona. The study reveals that YouTube generates moderate filter bubbles for most personas. However, the filter bubble effect is weak for personas who engaged in niche content, such as Conspiracy and QAnon channels. Surprisingly, all political personas, excluding the mainstream media persona, are recommended less videos from the mainstream media content category than an anonymous viewer with no personalisation. The study also shows that personalization has a larger influence on the home page rather than the videos recommended in the Up Next recommendations feed.

Understanding the Social Mechanism of Cancer Misinformation Spread on YouTube and Lessons Learned: Infodemiological Study.

A knowledge gap exists between the list of required actions and the action plan for countering cancer misinformation on social media. Little attention has been paid to a social media strategy for disseminating factual information while also disrupting misinformation on social media networks. The aim of this study was to, first, identify the spread structure of cancer misinformation on YouTube. We asked the question, "How do YouTube videos play an important role in spreading information about the self-administration of anthelmintics for dogs as a cancer medicine for humans?" Second, the study aimed to suggest an action strategy for disrupting misinformation diffusion on YouTube by exploiting the network logic of YouTube information flow and the recommendation system. We asked the question, "What would be a feasible and effective strategy to block cancer misinformation diffusion on YouTube?" The study used the YouTube case of the self-administration of anthelmintics for dogs as an alternative cancer medicine in South Korea. We gathered Korean YouTube videos about the self-administration of fenbendazole. Using the YouTube application programming interface for the query "fenbendazole," 702 videos from 227 channels were compiled. Then, videos with at least 50,000 views, uploaded between September 2019 and September 2020, were selected from the collection, resulting in 90 videos. Finally, 10 recommended videos for each of the 90 videos were compiled, totaling 573 videos. Social network visualization for the recommended videos was used to identify three intervention strategies for disrupting the YouTube misinformation network. The study found evidence of complex contagion by human and machine recommendation systems. By exposing stakeholders to multiple information sources on fenbendazole self-administration and by linking them through a recommendation algorithm, YouTube has become the perfect infrastructure for reinforcing the belief that fenbendazole can cure cancer, despite government warnings about the risks and dangers of self-administration. Health authorities should upload pertinent information through multiple channels and should exploit the existing YouTube recommendation algorithm to disrupt the misinformation network. Considering the viewing habits of patients and caregivers, the direct use of YouTube hospital channels is more effective than the indirect use of YouTube news media channels or government channels that report public announcements and statements. Reinforcing through multiple channels is the key.

Deradicalizing YouTube: Characterization, Detection, and Personalization of Religiously Intolerant Arabic Videos

Growing evidence suggests that YouTube's recommendation algorithm plays a role in online radicalization via surfacing extreme content. Radical Islamist groups, in particular, have been profiting from the global appeal of YouTube to disseminate hate and jihadist propaganda. In this quantitative, data-driven study, we investigate the prevalence of religiously intolerant Arabic YouTube videos, the tendency of the platform to recommend such videos, and how these recommendations are affected by demographics and watch history. Based on our deep learning classifier developed to detect hateful videos and a large-scale dataset of over 350K videos, we find that Arabic videos targeting religious minorities are particularly prevalent in search results (30%) and first-level recommendations (21%), and that 15% of overall captured recommendations point to hateful videos. Our personalized audit experiments suggest that gender and religious identity can substantially affect the extent of exposure to hateful content. Our results contribute vital insights into the phenomenon of online radicalization and facilitate curbing online harmful content.

Science by YouTube: an Analysis of YouTube’s Recommendations on the Climate Change Issue

Citizens increasingly use online platforms, and in particular YouTube, to get informed. This paper addresses the role of YouTube’s recommendation system in science communication to assess whether YouTube's recommendation system is creating filter bubbles on the climate change issue. It, therefore, contributes to the growing literature that finds that such filter bubbles actually do not exist. The paper shows that users are likely to get content with similar opinions, at least for a few recommended videos in a row. In addition, it was observed that the same promoters of climate denial appear over and over again in different recommendations. Given all that, YouTube can lead to the creation of filter bubbles where users do not have access to all relevant scientific information.

“It Is Just a Flu”: Assessing the Effect of Watch History on YouTube’s Pseudoscientific Video Recommendations

The role played by YouTube's recommendation algorithm in unwittingly promoting misinformation and conspiracy theories is not entirely understood. Yet, this can have dire real-world consequences, especially when pseudoscientific content is promoted to users at critical times, such as the COVID-19 pandemic. In this paper, we set out to characterize and detect pseudoscientific misinformation on YouTube. We collect 6.6K videos related to COVID-19, the Flat Earth theory, as well as the anti-vaccination and anti-mask movements. Using crowdsourcing, we annotate them as pseudoscience, legitimate science, or irrelevant and train a deep learning classifier to detect pseudoscientific videos with an accuracy of 0.79. We quantify user exposure to this content on various parts of the platform and how this exposure changes based on the user's watch history. We find that YouTube suggests more pseudoscientific content regarding traditional pseudoscientific topics (e.g., flat earth, anti-vaccination) than for emerging ones (like COVID-19). At the same time, these recommendations are more common on the search results page than on a user's homepage or in the recommendation section when actively watching videos. Finally, we shed light on how a user's watch history substantially affects the type of recommended videos.

"How over is it?" Understanding the Incel Community on YouTube

YouTube is by far the largest host of user-generated video content worldwide. Alas, the platform has also come under fire for hosting inappropriate, toxic, and hateful content. One community that has often been linked to sharing and publishing hateful and misogynistic content are the Involuntary Celibates (Incels), a loosely defined movement ostensibly focusing on men's issues. In this paper, we set out to analyze the Incel community on YouTube by focusing on this community's evolution over the last decade and understanding whether YouTube's recommendation algorithm steers users towards Incel-related videos. We collect videos shared on Incel communities within Reddit and perform a data-driven characterization of the content posted on YouTube. Among other things, we find that the Incel community on YouTube is getting traction and that, during the last decade, the number of Incel-related videos and comments rose substantially. We also find that users have a 6.3% chance of being suggested an Incel-related video by YouTube's recommendation algorithm within five hops when starting from a non Incel-related video. Overall, our findings paint an alarming picture of online radicalization: not only Incel activity is increasing over time, but platforms may also play an active role in steering users towards such extreme content.

Examining the consumption of radical content on YouTube

Although it is under-studied relative to other social media platforms, YouTube is arguably the largest and most engaging online media consumption platform in the world. Recently, YouTube's scale has fueled concerns that YouTube users are being radicalized via a combination of biased recommendations and ostensibly apolitical "anti-woke" channels, both of which have been claimed to direct attention to radical political content. Here we test this hypothesis using a representative panel of more than 300,000 Americans and their individual-level browsing behavior, on and off YouTube, from January 2016 through December 2019. Using a labeled set of political news channels, we find that news consumption on YouTube is dominated by mainstream and largely centrist sources. Consumers of far-right content, while more engaged than average, represent a small and stable percentage of news consumers. However, consumption of "anti-woke" content, defined in terms of its opposition to progressive intellectual and political agendas, grew steadily in popularity and is correlated with consumption of far-right content off-platform. We find no evidence that engagement with far-right content is caused by YouTube recommendations systematically, nor do we find clear evidence that anti-woke channels serve as a gateway to the far right. Rather, consumption of political content on YouTube appears to reflect individual preferences that extend across the web as a whole.

When algorithms steer the news: YouTube recommendations during the 2019 national elections

What people see and read in the media is no longer only determined by journalists, but increasingly by algorithms. These algorithms select, sort, and prioritize our information. Automatic processes like YouTube’s recommendation algorithm influence our views of the world. An important democratic question is whether YouTube’s recommendation algorithm incidentally exposes the audience to political information after watching entertainment content and whether the algorithm creates a filter bubble by primarily recommending content with a similar political perspective. During the Danish parliamentary elections 2019, the recommendation algorithm was more likely to lead viewers away from news and public affairs than towards political content. After watching a video posted by political parties Venstre or Stram Kurs, the algorithm primarily recommends videos from these same parties, which could strengthen confirmation bias and reinforce political beliefs. For other parties, this was less the case. Little evidence was found that the recommendation algorithm leads viewers from mainstream content to extreme right content.