>> Grizzly murders on Facebook Live on Tuesday in a deserted hotel in Thailand. A self-shot video of a father killing his 11 month old daughter by hanging her with a rope from a rooftop. The content lingered on the site for 24 hours. Barely, two weeks ago in Cleveland, a man bragging live about murdering a 74-year-old, before posting the video of the killing itself online.
These are just the latest in a string of violent crimes appearing on the social media site. Now, competing to solve the policing problem, tech firms from Singapore to Finland raising to improve artificial intelligence. So it can automatically stop and block content before it goes viral.>> So it'll detect that is a person that will detect that is a baby it will detect the activities that the person is involved with the baby, it will detect the emotion of the person, it will detect any, the screaming sound of the baby
>> Graymatics is an artificial intelligence company in Singapore, using deep learning. A type of AI that mimics the way neurons work and interact in the brain. The teaching system feeds images of what might be a violent scene. Anything from weapons to hacking movements and blood. But the software can only identify examples it's been taught to watch out for, and before this, someone hanging a child from a building, may not have been one of them.
As a dozen more companies wrestle with the problem, many are focusing on this method. Facebook under pressure over how it monitors content and live feeds, and Google faces similar problems with its YouTube services. None so far claimed to have cracked the problem completely.