KidsTube: Detection, characterization and analysis of child unsafe content & promoters on YouTube

Abstract

Abstract-YouTube draws large number of users who contribute actively by uploading videos or commenting on existing videos. However, being a crowd sourced and large content pushed onto it, there is limited control over the content. This makes malicious users push content (videos and comments) which is inappropriate (unsafe), particularly when such content is placed around cartoon videos which are typically watched by kids. In this paper, we focus on presence of unsafe content for children and users who promote it. For detection of child unsafe content and its promoters, we perform two approaches, one based on supervised classification which uses an extensive set of videolevel, user-level and comment-level features and another based Convolutional Neural Network using video frames. Detection accuracy of 85.7% is achieved which can be leveraged to build a system to provide a safe YouTube experience for kids. Through detailed characterization studies, we are able to successfully conclude that unsafe content promoters are less popular and engage less as compared with other users. Finally, using a network of unsafe content promoters and other users based on their engagements (likes, subscription and playlist addition) and other factors, we find that unsafe content is present very close to safe content and unsafe content promoters form very close knit communities with other users, thereby further increasing the likelihood of a child getting getting exposed to unsafe content.

Detecting and Measuring the Exposure of Children and Adolescents to Inappropriate Comments in YouTube

“…Among other social media platforms, YouTube has been the subject of many studies since it is considered the most popular social media platform in the United States [21], and the second-largest search engine after Google worldwide [15]. Studying the appropriateness of contents being presented to children on YouTube was first considered, to the best of our knowledge by Kaushal et al [12] who studied kids-unsafe contents and promoters. The authors provided a framework for detecting unsafe contents using measures calculated on the video, user, and comment levels with an accuracy of 85.7%.…”

“…Recently, we [1] investigated topic-driven toxicity on news broadcasted on YouTube and found many toxic, hate and profanity posted on this platform, and their topic correlations. Researchers have spent enormous efforts understanding the age-appropriate experience of children and adolescents when using YouTube, and have shown that inappropriate contents-such as contents with sexual hints, abusive language, graphic nudity, child abuse, horror sounds, and scary scenes-are common, with promoters for such contents targeting this demographic [6,7,11]. Parents and custodians trust children-oriented YouTube channels, such as Nick Jr., Disney Jr., and PBS Kids, to present educational and entertaining material for their children even with no supervision.…”

Video spam comment features selection using machine learning techniques

“…Neural Network [2] KidsTube: Detection, Characterization Based on Table 3, the first author with research about Comment Spam Filtering in YouTube used ten comparison of classification algorithm which are Decision trees (CART), K -nearest neighbors (k -NN), Logistic regression (LR), Bernoulli Naïve Bayes (NB-B), Gaussian Naïve Bayes (NB-G), Multinomial Naïve Bayes (NB-M), Random Forest (RF), Support vector machines with linear kernel (SVM-L), Support vector machines with polynomial kernel (SVM-P) and Support vector machines with Gaussian kernel (SVM-R).…”

“…The availability of resources over the Internet and broadband connection enables the emergence of sophisticated new platforms. In this way, YouTube is a one well-known video content publishing platform with social networking features, such as support for posting text comments to provide interactions between producers (channel owners) and viewers [2].…”

Technology in Education: Looking Back From 2020

“…For example, social media usage could be linked with increased mental health issues, such as feelings of depression and loneliness (Hunt et al, 2018;Steers et al, 2014). Moreover, Youtube recommendation algorithms often show more and more extreme videos, which could lead to radicalization (Tufekci, 2018), and often recommend inappropriate videos to children (Kaushal et al, 2016;Papadamou et al, 2019). Reading comprehension seems to be on the decline and how much we read is changing (either increasing or decreasing depending on what one counts as reading) as a result of increasing reliance on digital media (Wolf, 2018).…”