Technology / 4 October 2019, 06:00am / Louis Fourie
CAPE TOWN – One of the major drivers of the Fourth Industrial Revolution (4IR) or the age of intelligentisation is the major advances in Artificial Intelligence (AI) that are supporting and even taking over from humans in many situations. AI is increasingly replacing humans where knowledge could be learned or the decision-making formula is known.
It is in particular the AI abilities of machine learning and deep learning that makes AI so powerful in numerous fields. Machine learning refers to the ability of computer systems to learn by itself and to adapt accordingly, allowing them to perform a specific task without explicit instructions. In the Business Report of last Friday I illustrated that AI even transforms the disciplines based on “human touch” such as social work and is used to predict successful youth influencers in an HIV campaign; match homeless people with the best-suited housing and most effective social interventions; and select vulnerable families and children in need of intervention.
In a related field, assistant professor Desmond Patton from Columbia University in the USA uses natural language processing algorithms to analyse gang violence – a technology that could possibly assist with the gang violence on the Cape Flats.
In his research professor Patton noticed that the growing violence in poor urban areas is related to the fact that most gang members are quite active on social media. His observation is in line with growing evidence worldwide that social media is able to catalyse and amplify hostile relationships among gangs who have longstanding feuds in the community. More than often these tensions between gangs resulted in serious injury and homicide.
In his initial research Patton discovered that social interactions of gang members play out on social media. The modus operandi of gang members have changed and many of them now have a public social identity and conduct gang business on their smart phones. Community conflicts are no longer limited to face-to-face tough talk on the street. Gang members now communicate threats on popular social media platforms such as Facebook and Twitter – behaviour termed by professor Patton as “Internet banging.”
The online social identity is usually typified by overstated masculinity, toughness, willingness to engage in violence, and significant empty boasting typical of street culture. Unfortunately these messages of violence and retaliation often are the spark that ignites the smouldering undercurrent of potential violence between rival gangs. Research has shown that for violence to spread three elements are needed: (1) concentration in a specific geographic space; (2) the reciprocal character of the violence; and (3) escalations in assaultive violence.
Sufficient empirical evidence suggests that Internet banging eventually leads to serious injury and homicide in many urban communities. The virtual spaces used by gang members to establish their gang affiliation and boundaries can abruptly turn into threatening environments characterised by violence.
What makes social media so different and dangerous in this regard is its broad, real-time communication capabilities. One gang member can reach hundreds or even thousands of other people and gang members in a matter of seconds wherever they are. One Twitter message by a gang member can therefore trigger an avalanche of communication between rival groups who fiercely compete for reputation, territory (virtual and physical), markets, and the upper hand in violent encounters.
Professor Patton therefore started to investigate if a method could be developed to predict when online harassment would exacerbate tensions among rival gangs and would most likely spill out into the streets in the form of physical gang violence. If gang violence incidents could be predicted, it would be possible to intervene early enough and prevent the violence that takes so many (often innocent) lives in these troubled areas.
Professor Patton and his team thus conducted a sociolinguistic study to decode gang member language in two of America’s most violence-torn cities, Chicago and Baltimore. By using former gang-members he developed in both cases a computer-based, natural language processing algorithm that could automatically detect and decode the high-stress language and online provocations of gang-involved youth on Twitter and Facebook, which could be used to prevent future fatal or non-fatal firearm (offline) violence.
In 2019 professor Patton and colleagues published the details of their initial study in the International Journal of Bullying Prevention under the title “When Twitter fingers turn to trigger fingers: a qualitative study of social media-related gang violence.” The researchers identified six forms of social media communication that were interpreted as threatening by participants: venting (sharing of emotion), dissing (humiliating and degrading), calling out (instigating a violent response), direct threats, rap lyrics and posturing. They consequently developed a framework for understanding participant responses to tweets and the potential for violence as a consequence of such posts.
The team found that the interpretation of tweets is complex and is very dependent on the understanding of the local offline conditions and numerous other contextual factors, such as the local language; the local geographical environment and territory; how and why gang members use social media; and other nuanced and subtle cues indicating the legitimacy of the author, the intent of the message and the seriousness of the threat and likelihood of violence.
This USA study has important implications for the prevention of gang violence that is amplified by social media communication. The use of social media has grown exponentially over the past few years and has penetrated the lives of most individuals from all layers of life, including the lives of active gang members.
Although the language, environment and use of social media by gang members may differ in South Africa, it is certainly possible through a thorough analysis of the social media to determine the possibility of an eruption of physical violence between gangs. It has been shown that “dissing” stops short of inciting violence and that “calling out” pose a much greater risk for initiating violence especially when it involves specific references or challenges to individuals, locations and groups. The greatest risk of violence is indicated by “direct threats” where individuals, groups and locations are specifically identified together with an immanent threat of violence.
Social media and messages of violence, dominance and retaliation have a large audience and thus considerable impact on the perceptions, cultural values and actions of a large number of individual gang members. It is an important outlet for the frustrations of the youth in certain communities, unfortunately often with dire consequences.
The use of social media data for the study of gang violence worldwide and in South Africa has important implications for the prevention of violence and crime on the streets of certain suburbs. For example, the various features of social media, which include text, images, video, emojis, and hashtags, could be used for algorithmic identification of behaviours, themes, and conversations that indicate the intent to commit violence.
Furthermore, social media may also provide deep insights into root causes of violence in certain marginalised communities. If we want to solve the problem of the continuing gang violence on the Cape Flats, we will have to bring together computer scientists, social scientists, data scientists, and community groups to create inclusive and ethical AI systems that can thematically group social media posts regarding relevant conversations and events into categories. Particular attention should be paid to those, which seem to promote violence or cyber bullying.
I earnestly hope that law enforcement agencies, violence prevention organisations, and schools will consider using social media and artificial intelligence as tools for the prevention of gang-related violence in communities and cyber-bullying in schools. It may just provide the breakthrough we have been hoping for.
Professor Louis C H Fourie is a futurist and technology strategist. [email protected]