Child protection is clearly a deadly serious matter, and Tiscali is offering an anti-grooming tool to help scan the conversations children have over the Internet and alert parents to potentially inappropriate content. The announcement is aimed to coincide with the results of the Byron Review into gaming and online safety; if you remember, Shiny Shiny previously interviewed Dr. Tanya Byron herself about this.
It all sounds like a great idea, if occasionally a bit odd. The software, made by Crisp, bypasses website-locking, blocking and standard word filters to analyse the content of conversations, identifying inappropriate relationships. It was apparently independently tested by Cambridge University researchers and found to be 98.4% effective; this leads to some rather disturbing thoughts about how accurate researchers pretending to be paedophiles could actually be.
It also looks at "sexual content, punctuation, aggression levels and sentence length, typing speed and vocabulary" as well as learning as it goes, adding to its live database. I don't know about you, but I read that as checking for sexual punctuation, and am now trying to imagine what a suggestive comma looks like. Meanwhile, I'd love to know how you can track aggression levels given than anyone! pretending! to! be! down! with! the! kids! would! presumably! fill! their! writing! with! overenthusiastic! punctuation! AND UNNECESSARY CAPITALS anyway!!!!!!!!
The service is available for £3.50 a month or £42 a year. If it actually works, it's a very small price to pay to ensure children's safety. But checking who they're talking to, what sites they're visiting and how long they spend on the computer is free.
Into fitness and health gadgets? Check out our new site, Connected Health
Check out the best iPhone 4 accessories here ,
Hi, Thanks for the article.
One thing I need to point out. No researchers either here or at Cambridge went online and posed as paedophiles, I am not sure what gave that impressions. But it would never happen, it would also be illegal.
The engine was tested with a mixture of real grooming conversations and suspect conversations. Some of those from individuals who are now in Prison because of their actions.
In all cases it is tested with real data, to be an accurate scientific test ,nothing else could be uses.
However it is a sad fact that nobody would need to pretend to be a paedophile, there are at any one time 50,000 of them online and 1 in 4 children will be solicited for sex online every year.
Regards
Peter Maude
Crispthinking
Sorry, should have made it clear that it was more the way that the testing was described that made me imagine that than actually thinking it was the case.
Thank you for commenting; the statistics are indeed frightening.