Think before posting videos of kids as AI may flag them as child abuse

23Dec - by aiuniverse - 0 - In Data Mining

Source: newindianexpress.com

THIRUVANANTHAPURAM: Next time when you are about to post the seemingly innocent videos and pictures of your kids on internet, think twice. It’s just not that the paedophiles would misuse the videos, but those videos will create hindrance to the law enforcement agencies as well who work overtime to detect abuse videos. 

As the abuse videos are detected by artificial intelligence-powered bots, the videos of kids posted without any sinister intention will also be read by the system as child-abuse videos, making it difficult for the agencies to segregate them. For Kerala police, the data on child abuse materials are provided by the Interpol and the Internet Crime Against Children-Child Online Protective Services (ICACCOPS), developed by the US Department of Justice. 

The Interpol prepares the data pertaining to particular states using their data mining artificial intelligence-powered bots. The data concerned is then distributed among the respective regions. However, the agencies have a difficult task of separating wheat from the chaff due to the presence of large number of videos that have actually got nothing to do with child abuse. The Interpol data is provided to the state police from Delhi once or twice a month. 

Each report usually will have mention of 100 to 150 cases of abusive materials available on torrent sites as well as peer-to-peer networks. Similarly, the ICACCOPS- that’s a portal which can be accessed any time- also provides enormous data on child abuse materials pertaining to the state.

Facebook Comments