1,100 schools in the United States crawl students' social networks. They are looking for signs of depression and addiction to violence.

When potential problems are detected, depending on their severity, the school administration, parents and, in

extreme case law enforcement officersauthorities are notified of this. In the fall of 2017, Bark conducted the first pilot test involving twenty-five schools. “We found several rather alarming problems, including the threat of a bomb explosion and a shootout at the school,” notes the director of the parent committee of the company, Titania Jordan. After that, they offered to introduce large-scale monitoring.

Bark's product is free for US schools.for an unlimited time. The company notes that it can allow it because of the money it earns on the full version for parents. The cost of the parent product is $ 9 per month and includes monitoring more than twenty-five social networks, from Twitter and Instagram to Snapchat and YouTube.

None of the companies claim thatcan detect suspicious behavior every time - sometimes they get false positive results. However, when “suspicious content” is detected, the program sends a message to parents and schools with recommendations on how to solve this problem.

Now programs are used in more than 1100schools, it is about 2.6 million children. If materials that can be called “extremely serious” —for example, kidnapping a child or the threat of school shooting — are discovered, the matter is referred to the FBI.

YouTube headquarters attacked. Four people injured

According to Jordan, the app sends from 35 to55 thousand warnings every day. But since Bark launched a school product in February 2018, the FBI has received reports of sixteen likely school shootouts.

"Studies have shown that children communicate betweenbefore committing the act of violence. If we don’t have the means to hear these calls for help, the children are in danger, ”said Bill McCullough, vice president of sales for Gaggle.

McCullough added that their company does not relymachine learning only to identify threats. If, as a result of scanning letters and documents issued by the school, a child is found in a crisis situation, qualified security specialists analyze their content and check whether the threat is legitimate, and then determine the degree of its seriousness.