A woman takes a photograph with a camera whilst standing against an illuminated wall bearing Snapchat Inc.’s logo in this arranged photograph in London. (Photographer: Chris Ratcliffe/Bloomberg)

Tech Companies Identify, Remove 40,000 Terrorist Videos, Images

(Bloomberg) -- Big technology companies have added the digital signatures of 40,000 terrorist videos and images to a shared database as they seek to keep extremist content off their platforms.

Facebook Inc., Google’s YouTube, which is owned by Alphabet Inc., Microsoft Corp., and Twitter Inc. revealed the numbers in a joint blog post Monday.

The four big social media companies, which are part of a group called the Global Internet Forum to Counter Terrorism, announced one year ago that they would begin sharing digital fingerprints – known as hashes – of videos they removed from their platforms for terrorism.

Under the initiative, if a company removes a piece of content from its network for violating policies around terrorism, it is logged in the shared database. Then, if someone tries to post the same content to one of the other participating social networks, the content is automatically flagged for review – usually by a human analyst – and possible removal.

Technology companies have been under increasing pressure from Western politicians to do more to tackle terrorist propaganda and recruitment online. British Prime Minister Theresa May has been particularly active in accusing tech companies of not doing enough to keep extremists off their platforms and has called for international regulation to force the companies to do more or face substantial penalties.

The companies, for their part, have recently been highlighting their progress in using artificial intelligence and tools like the shared hash database to identify and remove terrorist content faster.  Last week, Facebook officials said in a blog post that their automated systems spotted 99 percent of the ISIS and Al Qaeda-related content the social network removes before the offending posts were flagged by users – and that in some cases the system was able to prevent the content from ever being published in the first place.

Twitter said in September that it also removed 99 percent of terrorist posts without relying on user complaints. YouTube said that, as of November, its algorithms automatically spotted about 83 percent of the terrorist-related content that the company removes.

As a result, terrorist groups have increasingly shifted their propaganda and recruiting efforts to newer, smaller social media platforms.

The four big tech companies said Ask.fm, Cloudinary, Instagram, which is also owned by Facebook, Justpaste.it, LinkedIn, which is owned by Microsoft, Oath, and Snap Inc. were going to begin participating in the shared terrorist content hash database.

Notably, however, Telegram, which has become a popular forum for terrorist groups such as ISIS to use to spread propaganda and reach out to potential recruits, is not part of the project.

“We recognize that our work is far from done, but we are confident that we are heading in the right direction,” the four companies said in their joint blog post.


To contact the author of this story: Jeremy Kahn in London at jkahn21@bloomberg.net.

©2017 Bloomberg L.P.