Will machines be able to detect cyberbullying in the future?

Written by Lizzie Learn On 27th Feb 2017

ith the advancement in technology and the ever-growing rate of “smart things” one has to wonder if Artificial Intelligence would be put to good use in day-to-day activity and life. One area where this may be of particular use is bullying in schools, and in particular - cyberbullying. The rates of bullying incidents are increasing worldwide, and in the case of cyberbullying the situation is even more critical.

Despite government policies and schools’ active approach in tackling bullying, it seems that there are numerous cases that go under the radar. What is more, children of this generation are rightfully considered “truly digital” being born after the Internet was invented. Their Internet usage habits are quite different from their parent’s and this has been a cause of concern in recent years.

Will machines be able to help?

You may not know it but even at this moment applications to monitor usage and set parental controls on all devices a child would use are already being developed and used by progressive parents. However, it seems they are only limited to observing children’s behaviour online and are not advanced enough to take autonomous actions and be proactive in warning parents when their child’s online interactions with other Internet users are suspicious. But given the recent developments in the Computer Science field and Artificial Intelligence in particular, there seems to be some hope that it may be precisely the non-emotional machines that will ultimately help us deal with such highly emotional issues.

How would a machine know if cyberbullying is occurring?

Artificial Intelligence programs process data and store it within their memory in special form - semantic networks. What this effectively means is that a program sorts the information it is processing and tries to make logical connections based on learning algorithms.

These learning algorithms are becoming so advanced that they are able to learn on their own without the need of explanation from a human side.

Take for example Google’s DeepMind which is an experimental program designed to understand the rules of board games given only the board and minimal input. Recently, DeepMind learned how to play the game of Go on its own and succeeded in playing the game against a human and winning. The same way, in the future we can have helpful programs that will be able to learn the concept of bullying and cyberbullying from manuals and other documents that we provide.

When we use the program on a child’s device or computer, it could be actively listening and monitoring online usage just like programs that exist now. But the difference will be that behind the scenes, the program will analyse the behaviour and compare it to what it already knows from bullying manuals, it will be able to make an accurate prediction whether an innocent conversation online will turn into a harmful one.

Utopia or Dystopia?

This all may sound a bit Orwellian but we need to be prepared for what may be a normal situation in the future. Taking an active approach towards developing and training such technology and installing programs that will passively listen to the conversation that our children are having even from the tiny microphone on their smartphone, could be the only real way to prevent bullying before it occurs. It is within the scope of technology to develop such systems within the foreseeable future although the question stays - would we want to use them? Would we trade the privacy of our child in return for them being practically bully-proof? When speculating about a future scenario of such kind, one must always ask the question:

What could possibly go wrong?

And in the case of a smart program that acts as a bullying vigilante 24/7, the answer is a lot. Perhaps even in ways that are not immediately apparent.

Privacy aside, when there isn’t a need for humans to be vigilant of bullying cases because they know there is always a machine that will do it for them, chances are there will not be a real need to educate students and teachers about it either.

Despite the fact that bullying is undeniably evil, when learning about bullying the child is actually developing more than just factual behavioural knowledge, they develop emotional sensitivity and a moral compass which are far more important lessons for the future. Whether you like the idea of smart programs helping us humans prevent bullying of children in schools and online, you probably agree that even thinking about such a potential scenario only highlights the importance of discussing the topic. After all, latest statistics show that around half of young children experience bullying in school, one way or another. Until our technology advances enough in order to help with the issue, we must continue the active discussions and attempt to solve this problem on our own.