Η Google άρχισε να διαθέτει νέα πιο ισχυρά εργαλεία AI στους διαφημιστές της. Η κίνηση ήρθε όταν κορυφαία εμπορικά σήματα ανέφεραν στην Google ότι είδαν τις advertisements them to be featured in propaganda videos via YouTube.
Admittedly, advertisers do not want to fund extremism, albeit inadvertently, and several of the companies withdrew their ads.
So Google has reportedly developed a new improved algorithm to combat such incidents. According to Bloomberg, the company is now using an upgraded form of machine learning to identify inappropriate content, five times more than it did before.
The new algorithm it has reportedly been running for the past two weeks already.
YouTube's huge video library is very difficult to control, at least by humans. So it was inevitable that some of the audit work would have to be outsourced to algorithms AI.
Google's General Manager for Business Development, Philipp Schindler, told Bloomberg that human eyes have contributed to labeling but that they can not do everything.
"The problem cannot be solved by man and should not be solved by man."
