I’m not a computer programmer, nor do I spend any time chatting on forums, but I was involved in internet publishing for many years and have seen the ease with which a few ordinary people (known as trolls) can hijack any constructive online dialogue and drag it down into an underworld of bitterness and mean-spirited vitriol.

The major social media companies do try to control this to an extent, but not always, and some of the stories of persistent bullying and persecution (of young people and adults alike) are shocking.

It’s all our own fault of course!!! They give us the tools to manage and control permissions and access rights, but they don’t make it easy. As an example, anyone who has finally given in and signed up to Facebook Messenger will probably feel a little uncomfortable about all of the permissions they’ve handed over.

Most of us don’t have the time to study the rules and learn how to manage all of our access rules in detail. Random postings still appear on my news feeds which have nothing to do with my friends or their interests – I wouldn’t have a clue how to block them.

The point is that, whether via over-enthusiastic friends or loopholes in our permissions, most social media users are exposed to externally-generated postings. These may or may not be harmless and the way they affect us is changing.

What exactly is a Troll?

It used to be that a troll was a private individual. An active forum or chat room user, often small-minded and empowered by anonymity and the protection of a keyboard. They would attack individuals, create conflict after conflict and, however many times they were locked out by forum administrators they would find a new identity and get back in. To troll was to attack someone or something online.

Those people still exist, but there are new dimensions to modern trolling. A growing number of troll ‘armies’ are appearing around the world. Some are owned and managed by governments or political parties, while some are commercially motivated mercenary groups, guns for hire in ever-evolving battlefields.

Unlike individuals, troll armies are trained and co-ordinated and, even worse, their employers keep them moderated and disciplined which makes them hard to identify and block. Each will control numerous social media accounts and will make multiple postings daily.

For example, with probably the largest and most-established troll army, the Chinese state pays a network of over two million people to post almost half a billion pro-government comments every year.

The strategy isn’t always one of direct praise. An outrageous or shocking claim made at the right time (by enough people) can deflect anger from a political scandal or a poorly-managed natural disaster. Or a tidal wave of warm, positive comments (cheerleading posts) can feed enough white noise to swamp legitimate protests.

And Bots?

As we all know, people can be a pain to deal with. They don’t always do what they’re told, they gossip, they change sides and they’re sometimes overcome with ‘totally unreasonable’ ethical objections.

In today’s world where computers and robots are set to take over more and more human tasks, why bother with a human troll?

The answer is fairly obvious – algorithms and programs can’t quite replicate the nuances of a real person, so most troll armies are obliged to manage with a combination of human and robot troops (which are know as “Bots”).

Bots are used across the internet in many positive ways; to keep things tidy, to improve bad spelling, to remove repetition etc. Anything that involves simple, clear repetitive tasks.

For the troll army commanders, the advantages of bots are that they cost less, don’t argue, don’t sleep and can do much more, much more quickly. The disadvantages are that they are more predictable and easier to identify and block automatically. It also requires more technical skill to manage bots.

What does this mean in the real world?

In future blogs I will look at some recent examples of social media manipulation including Brexit and the US and French elections, but before doing that, we should question whether it is really that easy to influence reasonable people using social media. Are we really that gullible?

Luckily, the nice people at Facebook have carried out some research which helps to answer that question? More on that to follow next time…

But before that, I think that it is worth taking a moment to imagine how advances in Artificial Intelligence (AI) will affect this landscape. How long will it be before AI bots are indistinguishable from human trolls? And then how long before user-friendly apps will appear which will allow any of us to build our own troll army?

What happens then?