The UK government had its own AI built to block extremist video
New software, intended for use by small tech firms, can pick out ISIS propaganda with reasonable accuracy.
The news: Developed by London-based data science firm ASI on behalf of the British government, the $800,000 AI is trained on thousands of hours of video to spot extremist content. It can identify 94 percent of ISIS propaganda with 99.99 percent accuracy, and is designed to be used to block the upload of such material.
Who it's for: It will be offered to small tech firms that can’t afford to develop such systems. The likes of Facebook and YouTube already have similar algorithms, though the British government has been critical of how widely and quickly they have been put to use by tech giants.
AI by law? The UK’s home secretary tells the BBC that use of the new AI could potentially become a legal requirement.
But: Because the AI is trained on historical data, extremists will likely be able to develop new ways to circumvent its watchful eye.
Deep Dive
Tech policy
2022’s seismic shift in US tech policy will change how we innovate
Three bills investing hundreds of billions into technological development could change the way we think about government’s role in growing prosperity.
Mass-market military drones: 10 Breakthrough Technologies 2023
Turkish-made aircraft like the TB2 have dramatically expanded the role of drones in warfare.
We’re witnessing the brain death of Twitter
An analysis of Musk’s tweets shows him at the center of conversations once kept on the fringes of Twitter.
Abortion pills via telemedicine: 10 Breakthrough Technologies 2023
Medication abortion has become increasingly common, but the US Supreme Court’s decision to overturn Roe v. Wade brought a new sense of urgency.
Stay connected
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.