Twitter takes further steps to promote 'healthy conversation'

Social network Twitter has unveiled a new approach to tackling the problem of disruptive trolling by utilising behavioural tracking tools.

In addition to the usage policies, review processes and machine learning tools it currently employs, Twitter will be integrating new behavioural signals into how Tweets are presented, in order to better identify users who regularly engage in troll-like behaviour.

The new system will look for signals such as accounts that have not confirmed email addresses, people signing up for multiple accounts simultaneously, or individuals repeatedly sending messages to accounts that do not follow them.

Content from these accounts will not be removed if they do not violate Twitter's terms of use, but these posts may end up being automatically hidden or excluded from searches, unless the user specifically chooses to view them.

Early testing of this approach have resulted in a four per cent drop in abuse reports from searches and eight per cent fewer abuse reports from conversations, which Twitter cites as evidence that this method can help to promote "healthy conversation".