When I met up with Dave King, chief executive of Digitalis Reputation, to discuss his research into how extremists are exploiting the Internet, I could not have imagined that just two days later three terrorists would launch an attack on London Bridge.
While prime minister Theresa May talked about tackling ‘safe spaces online’, King has identified another major issue: put simply, ISIS is a ‘great’ content marketer. (Hold on! Don’t splutter out your coffee in horror: I used inverted commas.) But his exhaustive research does reveal that the terrorist group uses SEO techniques, like any major brand, to ensure that its messages rise to the top of the search engine rankings for those who might tentatively seek some answers about its cause.
It is not building Facebook groups and Twitter accounts in the hope that potential followers might seek them out, but is instead aggressively pushing its messages to the top when its algorithm finds a web user showing an interest.
The report A War of Keywords, which was commissioned by the Centre on Religion and Geopolitics, an international think tank that analyses the interaction of religion, conflict and geopolitics, draws on Digitalis Reputation’s expertise – its technology is able to displace legacy issues for businesses from search engine listings. In short, it knows how the game is played.
As King explains: ‘The Centre explained to us the stereotypical journey of a teenage boy or girl in the west, who is not in any way radicalised but who is possibly infuriated by society and highly vulnerable, who goes online looking for information around themes that have interested them in the media, like caliphate.
‘And the fascinating hypothesis, which triggered the research, about the recruitment of interested moderates was that they weren’t suddenly following a radical preacher on social media, as governments might assume, they were Googling the phrases, like caliphate. And the big questions we then asked were, are ISIS and other terrorist groups researching their markets, are they engaging in SEO to promote their assets and are they becoming increasingly sophisticated in this area?’
Using SEO to find potential recruits
Indeed, Digitalis’ initial research identified half a million monthly searches on Google in the UK around phrases of interest. ‘ISIS know that the start of the journey of a vulnerable interested moderate is a search on a search engine for an innocent phrase. These aren’t people looking for beheadings,’ says King. ‘Much like anyone in marketing or recruitment is increasingly trying to promote their site to rank highly, in the very same way ISIS and other groups are absolutely promoting their websites to rank at the top of the Google searches.’
As King says, this means that the innocent teenager looking for information will instead find violent content on websites linked to terror groups. The counter narrative, which is produced by dozens of respectful Muslim groups, just does not rank. ‘The level of sophistication being applied to this grooming and recruitment is high. Often what is presented on page one of Google is not seriously radical, but it gets users onto these sites which are controlled by terror groups, and then there is almost a filtering process where they will gradually be exposed to more and more radical content. And some will drop away,’ says King.
‘I’d say it is as sophisticated as some of the very best marketing online, where brands look for particular types of users, and will only pay to reach them. And then the brand will filter that group down to reach the very best customers.’
The terror groups have also learned how to exploit Google’s ranking algorithm, which reportedly takes account of 200 different factors, but is particularly favourable towards link-building. If more people link to or reference an article on social media, it must be more interesting than an article without any. Thus, an online library, which contains literature and sermons from prominent jihadists, has almost 380,000 links into it from other sites. Other sites had ‘millions of links, likes and tweets into a particular page’, says King. Some were ‘natural’ links while others were specifically created to boost SEO. ‘There is absolute deliberate promotion being invested in, either through money or man hours.’
He admits to having been ‘shocked’ at the findings. ‘It is not just that extreme content is present when you search for these innocent phrases, but the fact that it is often dominant on page one of a Google search.’
Tackling the radical content
So where is the balanced content, that offers a counter narrative to that posited by the terror groups? It is there but it is just not visible. In only 11 per cent of searches conducted by Digitalis did counter-narrative content outperform that generated by terror groups. It is not because counter-narrative does not exist, nor that Muslim groups, NGOs and governments are not investing heavily in its creation. (The research found that 91 per cent of counter-narratives were Muslim-led efforts, indicating their online fight against this extremist ideology.) It is simply not working. The counter-narrative is displayed on websites that simply do not rank highly. ‘When we searched for counter-narrative we found lots,’ says King. ‘When we searched for counter-narrative around innocent phrases, we found none.’
In part, this is due to the fragmented nature of the approach. Each group, and factions within those groups, produce their own content. There is no attempt to work together to create a powerful response. ‘This counter-narrative needs help to get it ranking where it should be,’ says King. ‘We don’t see any investment in SEO on that side. Terrorist content is dominant today but in two years’ time it will be totally dominant. We’re behind. We can catch up but it requires significant investment.’
The findings are at odds with the Government’s focus of its counter-terrorist recruitment initiatives on social media. Does this mean they are looking in the wrong place, I ask. ‘For some reason, this is a whole area that is being ignored, partly because of a justifiable focus on social media. But we are missing the seed of the journey of the grooming and recruitment of our youth. Where does radicalisation start? It starts online with a search engine, like so many research journeys do, and that’s the bit that has seen no focus today.’
But King is not blaming search engine providers. ‘I have an immense amount of sympathy for Google and others,’ he says. ‘The Government and media are quick to bash them and say ‘Why didn’t they spot this content?’ The reality is that Google is trying to index an enormous amount of information, and whilst it will consider the semantic and contextual relevance of one page versus another, it is not really built to monitor and understand the nuances of certain types of content. Whilst they have made great strides to combat child pornography and abuse and whilst there are great technologies around to identify images that contain a lot of nudity, say, it is very difficult to create automated, or even human, systems that can differentiate the huge grey area between outrageous extreme and moderate.
‘I have sympathy with Google not wanting to be the arbiter in that area, and nor do I think we want a commercial operation to be so. There are a lot of nuance and grey areas.’ King points to the criticism that Facebook came under after it emerged that one of Fusilier Lee Rigby’s murderers had posted two years previously that he was going to kill someone. Why hadn’t the platform flagged that, people asked. Yet Facebook has more than 5.5 billion users who create five billion Likes per day. Its ability to distinguish a single credible ‘kill’ threat is minimal, to say the least.
‘I have to sit back pragmatically and ask how many people say they are going to kill someone in jest?’ says King. ‘These are brilliant organisations built for a specific purpose. We shouldn’t assume they are able to work miracles.’
The solution to overwhelm the terrorists’ content is, quite simply says King, to beat them at their own game. This would require the private sector, religious groups, NGOs and governments to work together. ‘It does require investment and it probably requires cross-party collaboration,’ he says. ‘It doesn’t take a tremendous amount of sophistication on the part of terrorists to ‘win’ the [online] battle, if nobody else is taking part.’
King concludes: ‘If we want to cut this [terrorism] off at its roots, we need to go back to the start of the search journey.’