We can teach our machines to delete ISIS clips – Google
IT HAS been heavily criticised for its woeful inability to remove jihadi videos and other shocking posts from the internet.
But Google is finally claiming to have come up with a solution to crack down on vulgar content online – computer systems that can be ‘offended’ like humans.
The tech giant has long been condemned for its powerlessness over what is posted on its online platforms.
And after last month’s Westminster terror attack, the Daily Mail highlighted how easy it was to use Google to find terror handbooks encouraging jihadis to carry out similar car and knife attacks.
But appearing to bow to pressure for more blocks on offensive content, the company says it is building computers which can learn context – helping them to identify anything that should be removed.
Google hopes its systems will, for instance, eventually be able to discern the difference between a jihadi with a gun and a scene from an action film – and take down the inappropriate material.
Currently teams of humans are checking the systems to see if they are doing a good job, but ultimately the firm hopes the machines will take over.
Google now wants the computers which monitor content being uploaded through YouTube and its other channels to understand the nuances of what makes a video offensive. The company is giving its systems human-vetted examples of safe and unsafe content as a reference point. Each video is then broken down frame by frame and every image is analysed individually.
Google also listens to what is being said on videos, reads the description and looks for other potential signals that could make a video inappropriate.
It claims it has already flagged five times as many videos as inappropriate under the new system, though it would not give exact numbers. The Mail has previously called on the biggest web firms to invest far more in making the internet safe or face legal action. Khalid Masood murdered four and injured more than 50 in a rampage outside Parliament on March 22.
In the wake of the attack internet companies including Google, Facebook, Microsoft and Twitter caved to Home Secretary Amber Rudd’s demand to set up a database of extremist content.
And the internet search engine has also come under fire from major companies whose advertising has appeared alongside distasteful pages and videos uploaded by terrorists.
Philipp Schindler, Google’s chief business officer, told the New York Times that the company has been taking these issues ‘as seriously as we’ve ever taken a problem’.
He said Google had been in ‘emergency mode’ since the Mail highlighted the problem the day after the London attack.
Mr Schindler added: ‘Computers have a much harder time understanding context, and that’s why we’re actually using all of our latest and greatest machine learning abilities now to get a better feel for this’.
Nobody from Google was available for comment.