The Anonymous Widower

The Problems Of Twitter

We have had a lot of stories in the past few years about people being abused on Twitter and similar sites, such as this report about Stan Collimore.

I have been involved in a few studies about malicious calls in the past with BT and have a small amount of knowledge. For instance, those that abuse usually target someone specific like an ex partner or employer or they might do the opposite and tend to target a range of people.

But there is always a pattern, as people are creatures of habit!

I also did some work years ago with project management software to see, if it were possible to fill in the blanks on an activity, based on other activities in a project and the words in the description. Techniques of language recognition, weren’t that good in the 1980s, but I believe that it is now possible to make a better fist of it.

Twitter messages are a string of a few words and I believe that computers can now get the flavour of what is said with a bit of help.  After all Google often gives you what you want in a search, that is full of spelling and typing mistakes.

So let’s say you get an abusive Twitter message. By flagging it to Twitter, they should be able to automatically determine if the contents fits a typical abusive pattern, based on the words and the tweeting profile of the sender.

An automatic warning would then ensue if it was deemed necessary, followed by deletion of the account, if the abuser persisted.

I obviously don’t know the thoughts of Twitter, but I’m absolutely certain, that an automatic system could be developed based on technology that works well in other areas.

The trouble is, any system like this is against the American rules on Free Speech. But it probably would be acceptable to many of those who have suffered abuse.

January 24, 2014 - Posted by AnonW | Computing, News | Abuse, Twitter, United States