Sort by
Role of Artificial Intelligence in the Prevention of Online Child Sexual Abuse: A Systematic Review of Literature

Online Child sexual abuse (OCSA) is a major menace in the digitalized world. Every year, more than a billion children between the ages of 2–17 years are sexually abused. Despite the harsh reality of most incidents, they remain unreported. Not only that but activities on dark web platforms go unmoderated and it becomes very difficult and complex to trace the origin of abuse and exploitation of children. In cases of online abuse, the sources through which detection can be done are evaded. Although several government and non-government initiatives have been implemented worldwide to curb this social menace, their effectiveness and accuracy remain questionable. AI-based services if regulated and executed cautiously can be effective in diagnosing and preventing sexual abuse in children on virtual platforms. Existing literature implies that AI can be a potent armor to detect and predict child sexual abuse online. The ability of artificial intelligence to predict and stop sexual abuse in children is very promising. AI-based technologies that can aid in the identification and prevention of violence against children include mobile computing, the Internet of Things, chatbots, machine learning, pattern recognition, and cloud computing. Therefore, it is essential to examine the existing literature pertaining to the research area, to highlight the emergent need for more applied and evidence-based research in the area. Thus, this study aimed to identify interventions driven by artificial intelligence in preventing online child sexual abuse, its limitations, and future implications. We conducted an extensive systematic literature review to understand the trends and efficacy of AI-based services in preventing and tracing child sexual abuse. The selection of research studies was performed in accordance with the PRISMA standards. Relevant studies were extracted from databases such as ScienceDirect, Springer, IEEE, and MDPI. Articles were selected and screened based on inclusion and exclusion criteria. This study identified 35 papers that were strictly limited to the prospect of AI interventions for online child sexual abuse. The review helped in deducing 3 major themes, namely, current trends in the field of AI for preventing online child sexual abuse, algorithm evaluation (advantages and disadvantages of AI tools) in preventing OCSA, and recommendation for technique advancements of AI tools. However, there is scarce evidence that proves AI interventions are effective in solving online child sexual abuse issues, as shown in the paper, thereby encouraging more extant research to be conducted in the area.

Just Published
Intelligence Collection Disciplines—A Systematic Review

Intelligence collection is an integral part of the intelligence cycle. In fact, some authors declare that it is at the heart of the intelligence discipline. Intelligence collection is typically done by a variety of intelligence collection disciplines and is as old as the Bible. In the past, intelligence collection consisted mainly of human intelligence (HUMINT). However, as technologies evolved, so too did collection methods, and the number of collection disciplines, therefore, increased substantially. Some of these intelligence collection disciplines also underwent some significant modifications because of these technological advances. An example of this is Image Intelligence (IMINT) which was previously seen as a collection discipline on its own. IMINT is nowadays considered a subdiscipline under Geospatial Intelligence (GEOINT)—the addition of geographical information systems (GIS) in the 1980s is one of the reasons for this change. These and many other changes resulted in many authors not agreeing on the main disciplines (and subdisciplines) in the intelligence collection domain. Furthermore, different organizations may only perform certain intelligence collection tasks and therefore only consider a certain spectrum of the intelligence collection domain. In 2021, the South African National Defence Force started a new degree programme in Defence Intelligence Studies under the auspices of the Faculty of Military Science, Stellenbosch University. It was, therefore, necessary to first establish what is globally considered the main intelligence collection disciplines and subdisciplines and secondly, which of these must be included when presenting intelligence collection as part of the degree programme in South Africa. The research entailed a two-phased approach, the first part entailed the PRISMA model to find relevant material that was analyzed with ATLAS.ti software during the second phase. The research is interesting since it suggests an expansion of the traditional list of intelligence collection disciplines by adding newer intelligence collection disciplines such as Social Media Intelligence (SOCMINT) and Cyber Intelligence (CYBINT). These additions can also be applied to other educational institutions offering intelligence studies elsewhere in the world.