UK Combats ISIS Videos With AI Technology That Detects Propaganda

Britain’s Home Office unveiled a tool Tuesday that algorithmically detects ISIS propaganda videos on small video hosting sites.

The technology, developed by ASI data science, is designed to detect and remove videos that have been created by ISIS. Thousands of hours of ISIS video content was analyzed in order to “teach” the software's artificial intelligence what to look out for. The program looks for certain cues, but much of the proprietary information wasn't released for security reasons.

The British government said that the technology was developed in order to prevent smaller video content publishers without large budgets from inadvertently spreading ISIS content. The project cost $834,000 and will be available to publishers for free. Smaller publishers cited include Vimeo, Telegra.ph and pCloud.

The software can detect up to 94 percent of ISIS content with a 99.995 percent accuracy, according to the company. It will flag non-ISIS videos .005 percent of the time.

“The purpose of these videos is to incite violence in our communities, recruit people to their cause, and attempt to spread fear in our society,” said Amber Rudd, home secretary in a statement. “We know that automatic technology like this can heavily disrupt the terrorists’ actions, as well as prevent people from ever being exploited to these horrific images.”

Rudd is in California this week meeting with Silicon Vally executives about preventing the spread of terrorist content online.

“Lone wolf terrorism is on the rise,” said ASI CEO Marc Warner in a promotional video. “Personally I was in Boston when the marathon bombings were happening and I was in London with the spate of recent attacks here.”

Warner said that as terror evolves, old tactics won’t work as well.

“It’s really hard for conventional intelligence agencies because one person being radicalized in a bedroom doesn’t show up under conventional surveillance techniques,” said Warner.