Artificial Intelligence could predict violent protests through Twitter analysis

Meanwhile, protests against WEF and TUS President Donald Trump are being held in Bern, Switzerland. (Reuters)

If someone wants to protest against an injustice or get people to support a cause, then Twitter provides the apt platform. In a democratic world, protests should be fine, but when the demonstrators hit the streets, there is always a chance of violence erupting between activists and authorities.

To avoid this, some researchers from the University of Southern California's Brain and Creativity Institute are using artificial intelligence to try and predict when protests will turn violent through Twitter analysis.

The researchers analysed the 18 million tweets posted during the Baltimore protests in 2015. Here is what had happened:

The research team tried to find an association between moral tweets and arrest rates. They found that the number of arrests was directly proportional to the number of moral tweets leading up to a protest. Also, whenever there were violent clashes between the activists and the authorities, it was usually preceded by tweets containing moral language or those related to issues that the posters could deem ‘Right’ or ‘Wrong’.

Another aspect they discovered was that when people moralise an issue that they are protesting, they are more likely to endorse violence. The Echo Chamber Effect was observed which simply means that if some people are confident that others in their social circle share their moral views, then they were more likely to consider violent attacks on their opponents.

By harnessing AI’s capabilities, it might soon be possible to predict violent protests by analysing tweets posted by protestors.