Can Artificial intelligence prevent Harassment and Cyberbullying?
Understanding The role of AI in preventing bullying, Harassment, and Discrimination
Cyberbullying is as dangerous and a huge concern as regular bullying. This social issue mostly targets children and teenagers. In the digital era, it’s becoming even more abhorrent, facilitated by the spread of technology. On the other hand, digital technologies like artificial intelligence (AI) can be a great help in preventing it.
Using Artificial Intelligence in preventing cyberbullying can curb the number of cases. Research by Sameer Hinduja and Justin W. Patchin reveals that around 5,000 surveyed middle and high school students in the United States alone, over a third have experienced cyberbullying. The age group of 12 and 20 are often targeted because of their appearance, performance, disabilities, religion, and other factors. To stop this, AI can be used as a technology for analyzing thousands of such cases, which is already happening in Japan. Let’s see how Artificial Intelligence is detecting, reducing and preventing cyberbullying cases.
Detecting and Decreasing Cyberbullying with Artificial Intelligence
In Japan, a 13-year-old boy took his own life after being bullied in 2011. In 2019, two former classmates of that boy were convicted and ordered to pay over $3000, 000 as compensation to the teen’s family.
The city had turned to Artificial Intelligence tools to prevent similar tragedies after such an incident.
Over 410,000 bullying cases were reported in Japan alone in 2017. These incidents are exacting a tragic toll. 10 of the 250 students who took their own lives had been bullied at school.
But while the access to technology is increasing the potential for bullying, it could also stump it out. AI tools are being used to analyse 9000 bullying cases, testing the grade and gender of the victims as well as the perpetrators. It aims to identify forms of bullying that can escalate. Moreover, computers powered by Artificial Intelligence are also being deployed to detect and deal with cases of harassment and bullying.
Gilles Jacobs, a language researcher at Ghent University in Belgium, and his team trained a machine learning (ML) algorithm to detect words and phrases linked with bullying on social media site AskFM that allows users to ask and answer questions. It managed to identify and block almost two-thirds of insults within almost 114,000 posts in English and was more accurate than a simple keyword search.
The exercise in identifying online bullying is far from merely academic. For example, a survey in 2017 found that 42% of teenage users have experienced bullying on the social media giant Facebook’s Instagram, the highest rate of all social media sites assessed in the study.
Instagram is now using Artificial Intelligence-powered text and image recognition to identify bullying in photos, videos, and captions. Wherein the company has been using a bullying filter to hide toxic comments since 2017, it began using machine learning in 2019 to detect attacks on users’ appearance or character, i.e., in split-screen photographs. It now also looks for threats against individuals that appear in photographs and captions.
Artificial Intelligence can not only help to stamp out bullying but also can save lives too. Around 3,000 people around the world take their own lives each day which indicates one death every 40 seconds. However, predicting if someone is at risk of suicide is notoriously difficult.