Inappropriate Content Research Articles

LLMGuard: Guarding against Unsafe LLM Behavior

Although the rise of Large Language Models (LLMs) in enterprise settings brings new opportunities and capabilities, it also brings challenges, such as the risk of generating inappropriate, biased, or misleading content that violates regulations and can have legal concerns. To alleviate this, we present "LLMGuard", a tool that monitors user interactions with an LLM application and flags content against specific behaviours or conversation topics. To do this robustly, LLMGuard employs an ensemble of detectors.

Efficient Toxic Content Detection by Bootstrapping and Distilling Large Language Models

Toxic content detection is crucial for online services to remove inappropriate content that violates community standards. To automate the detection process, prior works have proposed varieties of machine learning (ML) approaches to train Language Models (LMs) for toxic content detection. However, both their accuracy and transferability across datasets are limited. Recently, Large Language Models (LLMs) have shown promise in toxic content detection due to their superior zero-shot and few-shot in-context learning ability as well as broad transferability on ML tasks. However, efficiently designing prompts for LLMs remains challenging. Moreover, the high run-time cost of LLMs may hinder their deployments in production. To address these challenges, in this work, we propose BD-LLM, a novel and efficient approach to bootstrapping and distilling LLMs for toxic content detection. Specifically, we design a novel prompting method named Decision-Tree-of-Thought (DToT) to bootstrap LLMs' detection performance and extract high-quality rationales. DToT can automatically select more fine-grained context to re-prompt LLMs when their responses lack confidence. Additionally, we use the rationales extracted via DToT to fine-tune student LMs. Our experimental results on various datasets demonstrate that DToT can improve the accuracy of LLMs by up to 4.6%. Furthermore, student LMs fine-tuned with rationales extracted via DToT outperform baselines on all datasets with up to 16.9% accuracy improvement, while being more than 60x smaller than conventional LLMs. Finally, we observe that student LMs fine-tuned with rationales exhibit better cross-dataset transferability.

Statement of Removal

This Statement of Removal refers to: 
 Mastodon Content Warnings: Inappropriate Contents in a Microblogging Platform

ANALISIS BUKTI DIGITAL PADA APLIKASI DISCORD DESKTOP DENGAN MENGGUNAKAN FRAMEWORK DFRWS

A Currently, various communication platforms are available, and one of them is Discord. Discord is utilized by users to interact and share various types of content. However, Discord is often implicated in the distribution of inappropriate content that goes against Discord regulations. This can lead to cybercrime cases such as Child Sexual Abuse Material (CSAM). The main challenge faced is the recovery of digitally deleted data through these communication platforms. The objective of this research is to analyze digital evidence in the form of deleted text messages, images, and documents in the chat rooms of the desktop-based Discord application. The research method employed is the Live Forensic Technique, applying the DFRWS (Digital Forensics Research Workshop) method. The digital forensic analysis process is conducted using the digital forensic tool Magnet Axiom. The research results indicate that the accuracy of digital evidence analysis using Magnet Axiom reaches 94.11%. A total of 16 digital evidence items were successfully recovered out of the total 17 items available

Digital learning in childhood: Possibilities for pedagogical transformation in South Africa

Background: Digital technology (DT) has already changed history, and it is transforming childhood as more and more children go online around the world.Aim: The aim of this article was to critically explore the potential of young children’s digital learning (DL), while concurrently assessing the constraints.Setting: Early Childhood Care and Education (ECCE) stakeholders’ environments formed the setting for this study.Methods: The critical theory of technology emphasizes the need for increased democratic involvement in technical decisions. It also explores the impact of theories on technology users, affecting design, usage, and, consequently, outcomes. This study adopted an interpretative phenomenological approach to discover 10 South African ECCE stakeholders’ experiences of DL with young children.Results: Firstly, findings presented the affordances of DT. Secondly, the threats of connectivity, socio-economic factors, and the availability of inappropriate content were emphasised. Thirdly, there are risks and fears associated with DL, especially with young children. Lastly, the possibilities of DL, where pedagogical transformation can take place, were interrogated.Conclusion: In the ever-evolving world of DT, the pursuit of improved educational outcomes for young children remains a paramount concern. Proven practices that improve children’s learning through DT and effect change at a systemic level should be further interrogated.Contribution: This research contributes to the ongoing discourse surrounding DT in early childhood education by offering a critical examination of its challenges and opportunities. It provides insights for educators, policymakers and researchers on enhancing DL experiences for young children while considering the associated risks and benefits.

A Comparative Study of Responses to Retina Questions from Either Experts, Expert-Edited Large Language Models, or Expert-Edited Large Language Models Alone

ObjectiveTo assess the quality, empathy, and safety of expert edited large language model (LLM), human expert created and LLM responses to common retina patient questions DesignRandomized, masked multicenter study ParticipantsTwenty-one common retina patient questions were randomly assigned among 13 retina specialists. Each expert created a response (Expert) and then edited a LLM (ChatGPT-4)-generated response to that question (Expert+AI), timing themselves for both tasks. Five LLMs (ChatGPT-3.5, ChatGPT-4, Claude 2, Bing, Bard) also generated responses to each question. The original question along with anonymized and randomized Expert+AI, Expert and LLM responses were evaluated by the other experts who did not write an expert response to the question. Evaluators judged quality and empathy (very poor, poor, acceptable, good, or very good) along with safety metrics (incorrect information, likelihood to cause harm, extent of harm, and missing content). Main OutcomeMean quality and empathy score, proportion of responses with incorrect information, likelihood to cause harm, extent of harm, and missing content for each response type ResultsThere were 4008 total grades collected (2608 for quality and empathy; 1400 for safety metrics), with significant differences in both quality and empathy (p<0.001, p<0.001) between LLM, Expert and Expert+AI groups. For quality, Expert+AI (3.86 +/- 0.85) performed the best overall while GPT-3.5 (3.75 +/- 0.79) was the top performing LLM. For empathy, GPT-3.5 (3.75 +/- 0.69) had the highest mean score followed by Expert+AI (3.73 +/- 0.63). By mean score, Expert placed fourth out of seven for quality and sixth out of seven for empathy. For both quality (p<0.001) and empathy (p<0.001), expert-edited LLM responses performed better than expert-created responses. There were time savings for an expert-edited LLM response vs. expert-created response (p=0.02). ChatGPT-4 performed similar to Expert for Inappropriate Content (p=0.35), Missing Content (p=0.001), Extent of Possible Harm (p=0.356), and Likelihood of Possible Harm (p=0.129). Conclusions and RelevanceIn this randomized, masked, multicenter study, LLM responses were comparable to experts in terms of quality, empathy and safety metrics, warranting further exploration of their potential benefits in clinical settings.

Adolescence and Digital Well-being: Tools for Healthcare Professionals

Adolescence is a period characterized by significant physical, emotional, and social changes, including the acquisition of multiple skills. It is also a time when the risks of mental disorders, substance use, unwanted pregnancy, eating disorders, and negative effects of inappropriate social media use manifest greatly. Digital well-being implies the healthy use of technology, the protection of privacy and security online. Adolescents use social media to identify themselves, interact with others, entertain themselves, and seek information. However, risks are present, including access to inappropriate content, cyberbullying, problematic consumption, and fraud. The pediatric field must support adolescents in optimizing their online use. Implementing multisectoral strategies can minimize risks, promote digital literacy, and ensure equitable access to quality technological resources and telehealth, thereby promoting adolescent online well-being.

Famlogger: The development of keylogger tools for parental online monitoring

Recent advances in the Internet have brought about many benefits, but they have also created a number of major problems, particularly for children. One of the most pressing issues is the exposure of children to inappropriate content, such as cyberbullying, online grooming, phishing, and explicit material. This problem is exacerbated by the fact that children are often alone and unsupervised while using the Internet, making it difficult for parents to monitor their online activity. In order to address this issue, a developer has proposed the creation of FamLogger, a keylogger system that would allow parents to monitor their children's online activity from anywhere in the home. The system is designed to provide parents with real-time updates on their children's online activity, including the websites they visit, the apps they use, and the messages they send. To ensure that the system meets the needs of parents and children alike, a survey has been conducted to gather feedback and requirements for the system. The survey results are used to improve the quality of the system and make it more user-friendly for parents. FamLogger can help parents monitor their children's activity from anywhere in the home, providing them with peace of mind and giving them the ability to intervene if necessary.

‘I tried to take my phone off my daughter, and i got hit in the face’: a qualitative study of parents’ challenges with adolescents’ screen use and a toolbox of their tips

BackgroundConcerns about adolescent screen use are often expressed but poorly understood, particularly in terms of how parents are supposed to respond in ways that balance safety, care and developing independence and autonomy. This qualitative study investigated parental perceptions and concerns about screen use of adolescents aged 13 to 17. Current strategies to manage screen use and barriers to doing so were explored, and parents were asked to outline recommended interventions for better outcomes.MethodsInterviews and focus groups were held with 33 adults in Auckland, Aotearoa New Zealand (mean age 48 years) who were parenting adolescents (mean age 14 years). Interviews were transcribed verbatim, then inductive analysis and interpretation by the research team distilled the key ideas and illustrative quotes. A table of recommendations for a practical ‘toolbox’ was developed from these in-depth conversations.ResultsParents reported extensive use of screens by their adolescents, constantly throughout the day (and night, instead of sleeping). Four areas of specific concern included: (1) addict-like behavior, (2) exposure to harmful (and inane) content, (3) living in a virtual world, and (4) negative impacts on physical, mental, and cognitive wellbeing. To manage adolescent screen use, family rules and restrictions (on time and place) were common. Some used technical control via software or accessing the adolescent’s devices and/or accounts to check for inappropriate content (such as pornography). Communication about device use and self-regulation were important. Barriers to managing screen use included trying to avoid conflict with their child; difficulties with consistency or follow-through on rules; lack of technical knowledge; parental screen use that set a bad example; and device use needed for school or other purposes. Recommendations from parents are presented in a toolbox of tips and techniques they shared, and their ‘wish list’ for better access to practical, local, scientific information, examples of techniques that have worked for other families, tools for problematic behavior and risk (including how to begin conversations with adolescents about their concerns), and having schools and young people involved in developing interventions to build digital citizenship.ConclusionsRich, nuanced accounts from parents about adolescent screen use in their families and communities underpinned their practical ideas for more skillful responses to young people grappling with an addictive digital existence.

Legal Implications for Parties Who Display Inappropriate Content Through Social Media Platforms (Copyright Study)

Copyright is one of the rights included in protected intellectual property rights. One form of creation protected by the expert is the work of photography and cinematography. However, it is known that with the development of existing technology, works in the form of photography and cinematography are often used by other parties/people by showing and/or uploading the work to social media without permission from the creator himself. It is not uncommon for the photography and cinematography created to have economic value. Another problem arises when it turns out that the photography and cinematography created contain immoral elements. Therefore, a more in-depth study is needed regarding the legal implications of uploading photography and cinematography without the author’s permission and the limits given by the norms of decency in producing the creation itself. Based on the research results, it is known that the legal responsibility of the party who shows photography and cinematography without the author’s permission on social media platforms, namely if the action is to obtain economic/commercial benefits. It is proven that if the creator suffers economic losses, then the perpetrator can be charged compensation. This type of research is normative juridical/normative legal research, which uses secondary data by processing data derived from legal materials, namely primary, secondary, and tertiary legal materials. Study data were collected through documents and qualitative analysis.&#x0D; Keywords: copyright, intellectual property rights, information and electronic transaction law, photography, cinematography, social media

Decoding Desires: The Subconscious Motivations of Fan Fiction Authors

Fan-Fiction, originating from early science fiction, has become a popular culture, fuelled by the digital age and online communities. Fan-fiction is a creative technique in which fans actively connect with current media texts by creating their own stories, frequently reinventing and building on the established worlds and characters of their favourite source material. From a psychoanalytical perspective, it offers a secure environment in which young people may project their unconscious desires and thoughts onto the characters and settings they love. They can virtually experience events that could be controversial, banned, or otherwise deviate from society standards by inventing alternate narratives. Despite its easy accessibility, it can also be a source of objectification, inappropriate content, fluctuating writing quality and stereotypes, which could morbidly influence the psyche of young authors and readers. This study investigates fan-fiction from a psychoanalytic standpoint, concentrating on its diverse themes and attractiveness to the Fan-fiction community, aiming to understand its effect on its prosumers.

Utilizing Age‐Adaptive Deep Learning Approaches for Detecting Inappropriate Video Content

Utilizing Age‐Adaptive Deep Learning Approaches for Detecting Inappropriate Video Content

Exploring Gender Disparity in Geometry Learning Using Van Hiele’s Model: A Quasi-Experimental Study of Final Year Male and Female Students’ Performance

Children and teenagers are growing up in a society where social media and gadgets are commonplace. Although these technologies can be beneficial, they also pose risks, such as being exposed to inappropriate content. Parents must learn how to monitor and manage technology use in order to protect their children’s well-being. This study aims to explore the common parental strategy employ to mitigate the negative effects of technology on their children. The participants in this study were chosen using a narrative inquiry design and convenience sampling. After receiving permission, the researchers conducted an interview with parents in Opao, Mandaue City. The collected data will be analyzed thematically in order to correctly interpret themes and patterns. After analyzing the data, the researchers discovered that most parents would limit their children’s use of devices such as cell phones. Some parents, particularly those who are technologically savvy, would also provide appropriate guidance. The researchers have also found that communication is very important in making these strategies effective. Lastly, parents should strictly give proper guidance to avoid technology’s negative effects. The Department of Information and Communications Technology (DICT) and the National Bureau of Investigation (NBI) can work together to block access to pornography sites that are not suitable for minor children. School ICT coordinators should create educational videos suitable for children.

Unethical Behaviour in Distance Learning: Polish Student Perspectives

Background: During the COVID-19 pandemic, a transition occurred towards remote learning. This change posed various challenges. One critical issue was the emergence of unethical behaviour. This problem has the potential to detrimentally impact students' learning experiences.&#x0D; Research objectives: The study sets out to explore the array of unethical behaviors encountered by university students in distance learning. Additionally, it seeks to understand the disparity between students’ anticipations and the actual issues they faced.&#x0D; Research design and methods: The research employs a qualitative approach where data were collected from 24 students at the Krakow University of Economics through semi-structured interviews, focusing on their experiences with unethical behaviour.&#x0D; Results: The study unveils that students both anticipate and are apprehensive about occurrences of unethical behaviour in distance learning, including the sharing of inappropriate content. However, their actual experiences encompass different issues, such as unwarranted distractions. These distractions have a detrimental effect on the learning process.&#x0D; Conclusions: The study illuminates the incongruence between students' preconceived notions and their real-life encounters with unethical behaviour during online learning. It emphasizes the importance of promoting engagement and adhering to ethical standards. Furthermore, the study advocates for broadening the research framework to include the perspectives of educators, which is essential for a comprehensive understanding.

Interpersonal Communication Between Parents-Children in Increasing Children's Protection from Cyberbullying

In this digital era, children are very vulnerable to being exposed to the risks and dangers of technological developments, especially social media. Such as inappropriate content, online violence, cyber bullying, and many other impacts of technology and media. For this reason, it is important for parents to increase their children's protection from media exposure and the influence of technology in this digital era. Therefore, interpersonal communication between parents and children must be improved to protect children from the bad influences of media and technology. Parents need to improve good communication with their children so that they are willing to tell stories and share experiences about the problems they are facing, so that children feel comfortable and have a place to share their stories. Apart from that, parents also need to provide education regarding technology and its impacts and know the risks they will face, so that children can be more efective and wise in dealing with technological developments. The way that can be done is to provide adequate supervision and limit children's excessive use of media, so as to minimize the impact they will receive. This research article aims to determine interpersonal communication between parents and children so that it is effective and can increase protection for children from developments in digital technology such as cyberbullying. This research article uses qualitative methods with descriptive data processing techniques.

Preventive online safety education for teenagers

The rapid evolution of digital technology has brought both opportunities and risks, particularly for adolescents navigating the online world. This paper delves into the significance of preventive online safety education tailored for teenagers. It explores the current landscape of online threats, encompassing cyberbullying, privacy breaches, and exposure to inappropriate content, among others. Additionally, it evaluates existing educational approaches, emphasizing the need for a proactive, multifaceted strategy to equip teenagers with the necessary skills and knowledge to safely navigate the digital sphere. The proposed preventive education model integrates elements of digital literacy, critical thinking, and responsible online behavior. Furthermore, it discusses potential challenges in implementing such programs and suggests collaborative efforts among educators, parents, and policymakers to ensure comprehensive online safety education for teenagers, fostering a safer and more secure digital environment for the next generation.

Research on the Influence of Douyin on Teenager in China

In today's dynamic digital landscape, the influence of social media on adolescents has emerged as a central focus of research, education, and parenting. Douyin, a prominent short-form video platform, has become a key player in shaping teenagers' online experiences. This qualitative study undertakes a comprehensive exploration of the intricate relationship between Douyin and teenagers, shedding light on both its advantages and challenges. With a thorough review of pertinent literature, this research investigates Douyin's multifaceted impact on teenagers' psychological, social, and educational spheres. The study is guided by specific research objectives, including the examination of how Douyin fosters global connections and cross-cultural interactions, serves as a platform for promoting social awareness, contributes to skill development and informal learning, shapes teenagers' engagement with current events, and explores potential negative consequences such as addiction, physical harm, mental health issues, privacy concerns, and exposure to inappropriate content. The findings underscore the dual nature of Douyin's influence on teenagers, elucidating its role as a catalyst for global connections and social change, while also highlighting potential pitfalls related to addiction, safety, and privacy. As Douyin continues to play a pivotal role in the digital lives of teenagers, this research serves as a timely reminder of the importance of responsible use and guidance from educators, parents, and guardians. It underscores the need for a balanced approach to harnessing the platform's benefits while safeguarding against its potential drawbacks, emphasizing the broader significance of navigating the digital world with mindfulness and intentionality.

The Application of Short Videos in Oral Teaching for English Majors: Taking TikTok Short Videos as an Example

In recent years, there has been a growing demand for English major students to improve their oral communication skills. With the widespread use of mobile learning, students are now incorporating new technologies into their language learning. TikTok short videos have become an important tool for practicing oral communication due to their convenience, sharing capabilities, and multimedia features. This study takes students majoring in English from Sichuan International Studies University as the survey object and conducts an empirical study through questionnaires to explore their views on the application of oral learning short videos from TikTok to oral English teaching activities. The results demonstrate that TikTok short videos can increase student engagement and promote creativity in oral English teaching while students identify concerns around inappropriate content and potential distraction that may negatively impact learning outcomes.

Arabic Toxic Tweet Classification: Leveraging the AraBERT Model

Social media platforms have become the primary means of communication and information sharing, facilitating interactive exchanges among users. Unfortunately, these platforms also witness the dissemination of inappropriate and toxic content, including hate speech and insults. While significant efforts have been made to classify toxic content in the English language, the same level of attention has not been given to Arabic texts. This study addresses this gap by constructing a standardized Arabic dataset specifically designed for toxic tweet classification. The dataset is annotated automatically using Google’s Perspective API and the expertise of three native Arabic speakers and linguists. To evaluate the performance of different models, we conduct a series of experiments using seven models: long short-term memory (LSTM), bidirectional LSTM, a convolutional neural network, a gated recurrent unit (GRU), bidirectional GRU, multilingual bidirectional encoder representations from transformers, and AraBERT. Additionally, we employ word embedding techniques. Our experimental findings demonstrate that the fine-tuned AraBERT model surpasses the performance of other models, achieving an impressive accuracy of 0.9960. Notably, this accuracy value outperforms similar approaches reported in recent literature. This study represents a significant advancement in Arabic toxic tweet classification, shedding light on the importance of addressing toxicity in social media platforms while considering diverse languages and cultures.

Women and Girls in STEM Fields: How Gender Bias Impacts IndividualS Academic Decision

Female figures have experienced consistent stigmatization in a variety of domains; one of the most significant fields is STEM, which is perceived as a laborious and hardcore discipline. Women, who are generally correlated with characteristics such as delicate and tender, maintain a lower participation rate in both STEM subjects and jobs. In order to obtain a more profound knowledge behind this trend, this essay aims to investigate potential reasons by generalizing current issues into three sections: academic, familial and occupational. The major research method used will be the semi-structured interview, where some questions are pre-designed, and some are improvised. One sample question of the predetermined question is Describe the gender bias you have experienced. After an abundant collection of evidence, the essay comes along to the conclusion that schools, families and workplaces do hold crucial embodiment of gender bias in academic preferences. Deficits of the education system mainly include a shortage of models, inappropriate learning content and a discriminating academic atmosphere. The familial aspect displays its effects through parental disapproval and enforcement of changing majors. Meanwhile, STEM occupations also exist issues, including sexual harassment and maladjusted salary. Deterrence, as a result of these reasons, extensively blockades women from entering these domains. Thus viable blueprints of rectification, such as law establishment and additional support, are also brought up. Such measurements are firmly believed to be effective alleviations of the current status.