British Home Secretary, Amber Rudd, has today announced a new technology that will screen online video content for Daesh terrorist propaganda.
The technology has been developed in conjunction with ASI Data Science and, according to a video posted on ASI Data Science’s website, will rely on “subtle signals” to indicate the presence of extremist content.
The technology will “scan” video content before it’s uploaded and the “subtle signals” will be strong enough to distinguish between a Daesh propaganda video and other legitimate online videos, like news videos, which report on Daesh.
The software relies on a platform called Sherlock ML which ASI Data Science claims is the “world’s best platform for deploying data science models”. Sherlock ML utilises machine learning and has been ‘trained’ using 1,000 Daesh propaganda videos.
According to ASI Data Science, the technology has been developed solely for the purpose of screening Daesh propaganda videos, and no cross-application has been mentioned by the Home Office.
The development arrives in the midst of increasingly combative discussions between the world’s largest governments, online advertisers and mainstream media organisations and the world’s largest technology platforms they deem aren’t fulfilling their ‘responsibilities’.
Over the course of the last year, a number of changes have been made by the world’s largest technology platforms (including Facebook, Google and YouTube) to combat frequently cited complaints, including:
- YouTube has outlined plans to expand the size of the team tasked with policing its content to 10,000 people. YouTube has also stated it will apply warnings to content that contains “inflammatory or supremacist content” and issue disclaimers on content published by organisations which receive Governmental funding
- Facebook has made numerous changes to its news feed to tackle the proliferation of what is referred to as “fake news”. It has recently taken a decision to deprioritize national content in favour of local content and to make content from businesses, brands and media less prominent
- Google has introduced measures to allow brands to opt-in to appear next to “broader types of content if they choose”. It has also announced plans to introduce finer control over which sites advertisers want to exclude from their targeting.
In an official release published on the Gov.uk website, the Home Office states that larger corporations have reported on the significant difference technology has made to fighting extremist content online.
The release goes on to mention that that the technology will also be offered to smaller companies which are a target for Daesh propaganda and may not have the requisite resources to screen content being published on their platforms.
Some of the “smaller companies” highlighted include Vimeo, Telegra.ph and pCloud.
The Home Office also states that Daesh uses over 400 platforms to circulate extremist content online and research has previously found that the majority of links to Daesh content are disseminated within 2 hours of release.
Amber Rudd will this week meet with a number of large tech corporations and the Secretary of Homeland Security, Kirstjen Nielsen, to discuss how to tackle extremist content online.