It’s called “grooming” and it happens all the time. An adult poses as a child on the Internet for the sole purpose of tricking an actual child into having a conversation. Often, those conversations lead somewhere no child should go: Into the grasp of a pedophile.
In an effort to unmask these pretenders, Computer Scientists at Lancaster University are working on a software program that can determine a person’s age and gender by analyzing the language they use. While the software isn’t intended for children or their parents to use, the developers hope it will be useful to law enforcement agencies who are out there fighting the good fight against online sexual predators.
To assist in the development of the software, researchers recruited 350 children and teens from a school in the UK. Setting them up on the Internet, they monitored their online conversations with strangers. In addition to providing data for the software, these conversations underscored the need for such a program.
Of the 350 kids participating in the study, 4 out of 5 believed they were talking to a teen when in fact they were talking to an adult pretending to be a teen. Overall, only 18% of the kids guessed correctly, with the girls scoring somewhat better than the boys.
The software program, however, correctly determined whether a web chat was written by a child or an adult 47 out of 50 times.
In addition to determining the sex and age of an anonymous person online, the software has the potential to pick up on the ‘stylistic’ footprints of individuals, enabling law enforcement to follow pedophiles as they make their way around the Internet. Ultimately, the developers hope the software will automate the process of tracking online predators, freeing up police while at the same time helping them do a better job of catching these monsters before they strike in the real world.
More from this author: