Actions speak louder than words: the role of technology in combating terrorist content online
by Katie Passey
Internet platforms and emerging technologies are at the centre of the debate about how to beat online extremism. However, from narratives of AI successes to accusations of inactivity, it is unclear what work is being done and how successful it has been in the fight to prevent online radicalisation. UK Prime Minister Theresa May accused social media companies of providing platforms for the online radicalisation of violent extremists, she stated: ‘We cannot allow this ideology the safe space it needs to breed - yet that is precisely what the Internet and the big companies that provide Internet-based services provide’. Meanwhile, Facebook, Twitter and YouTube report statistical successes in their efforts to tackle online extremism: throughout 2017 Twitter suspended 574,109 accounts for violations related to the promotion of terrorism; between January and June 2018 over 75 percent of content flagged by YouTube’s automated systems was removed before it had received any views; and Facebook acted upon 1.9 million pieces of terrorist content between January and March 2018.
But there’s more to the story than these numbers suggest. Oxford Insights’ forthcoming report will illustrate the harsh reality that despite the impressive statistics and the evidenced capabilities of emerging technologies, more could be done by social media platforms to remove violating and illegal terrorist content online.
I interviewed Dr Hany Farid and Joshua Fisher-Birch from the Counter Extremism Project [CEP] on their recent reports - Okay Google, Show me Extremism and the eGlyph Web Crawler - both of which confirmed my suspicions that social media platforms are not doing as much as they could to combat online terrorist content. From hash sharing and image matching to conduct analysis and cluster profiling, there are a plethora of ways emerging technologies are being used to combat sinister content online; yet, when applied to online terrorist content these technologies are not being developed or used as aggressively as they could be. CEP found that online hashing systems are not working effectively to prevent re-uploads of known terrorist videos (with the same extremist content being repeatedly uploaded and deleted) and that a user searching for ISIS material on YouTube was more than three times as likely to encounter extremist material than counter-narratives. Many cases of known terrorist content from proscribed organisations are left on social media pages to inspire and recruit. Two recent examples found by CEP included: ISIS’ video ‘Graveyard of Enemies’, found on 29 August 2018 after two weeks with over 1,000 views, 93 likes and 80 shares, and ISIS’ video ‘You are not held responsible except for yourself’, found on 5 September 2018 after five days with 806 views, 90 likes and 13 shares.
The hindrance in removing non-state terrorist content does not derive from the technology, but alternatively the social media companies themselves. The action needed to effectively deny non-state terrorist organisations access to surface web platforms is not yet consistent with the business model and ideological principles of social media and communications providers. This is not to say that social media companies are deliberately encouraging terrorism, or deliberately making money from promoting it. Instead, platforms are struggling to find a balance between legitimate censorship and freedom of speech; they have tended to see themselves as communities that enable conversation and global connectivity, not the regulator of opinions. Twitter’s CEO Jack Dorsey this month argued that Twitter would ‘rather be judged by the impartiality of outcomes, and criticised when we fail’. A similar opinion was expressed by Facebook’s Chief Executive, Mark Zuckerberg, in March: ‘I feel fundamentally uncomfortable sitting here in California… making content policy decisions around the world… things like “where’s the line on hate speech?'“, I mean, who chose me to be the person that did that?’.
In terms of business models, being over-censorious, even of illegal content, can be bad for social media businesses as it could deter people from the platform if it is perceived that democratic principles are being compromised online. Social media platforms rely on advertising as their main source of revenue; fewer users on the platform clicking adverts means less revenue. As revealed by Cambridge Analytica’s psychographics, advertisers can entice users to remain clicking on platforms through tailored adverts constructed from a user’s extracted data activity across the Internet, to build a user psychological profile, and using that profile to show the user content that gets them clicking more and more. Cambridge Analytica’s former CEO Alexander Nix argued that: ‘if you know the personality of the people you are targeting, you can nuance your messaging to resonate more effectively with those key audience groups’. Typically, outrageous and controversial content gets users clicking the most because of its ability to evoke an emotional response that drives social sharing. Although this claim was disputed by Facebook’s Vice President of Public Policy Lord Richard Allan, who protested that ‘shocking content does not make us money, that is just a misunderstanding of how the system works’, it is not difficult to see how Facebook, and other platforms, might want to protect their high value content. This raises some uncomfortable questions about platforms’ ability to prevent online radicalisation and future efforts of governments to collaborate with them.
Emerging technologies have become useful in the fight against online radicalisation because of their speed, versatility and ability to process data more quickly than people can. However, until a strategy is in place that is mutually agreed between governments and social media companies on the ways in which non-state terrorist organisations are denied access to surface web platforms, the technology will only be as effective in the fight as social media companies dictate.
The next in this series of insights examines the final side of this relationship, asking: what role can the UK Government play in better combating terrorist content online?
We will be publishing a series of insights into the role of technology in non-state terrorist content online over the coming months, culminating in a full report. In the meantime, look out for part three in this three-part series. Missed part one? Click here. If you have any queries or questions, or would like to be involved with the report, please get in touch with Katie at firstname.lastname@example.org.