A couple of months ago there were reports that some ads on Youtube had appeared in offensive and/or violent content. As we exposed here, Youtube committed to improving its security procedures and guaranteeing the quality of the videos that are daily uploaded to the platform.
This resulted in the implementation of artificial intelligence that has proven to be highly capable of accurately identifying videos with unsafe content or favoring terrorism.
Among the highlights are:
– Increased speed and efficiency: 75% of extremist videos removed from YouTube were detected and deleted before even a single human user reported it.
– Precision: in many of the tests analyzed, the artificial intelligence platform was more accurate in their judgment than humans to determine if the video has inappropriate extremist content.
-Scope: With more than 400 hours of video uploaded to YouTube every minute, this new AI has been able to work twice as fast with double efficiency and precision when it comes to detecting and deleting extremist audiovisuals.
Despite this, Google recognizes the need for human intervention and has partnered with 15 additional organizations such as NGOs to provide expert knowledge on topics such as hate speech, radicalization and recruitment.
Higher standards: Videos that are not illegal but have been reported as potential violators of YouTube policies will be limited. These videos will not be able to monetize, will not be recommended and will not have tools like comments, likes and suggested videos.
Early intervention and counter-terrorism: A new method called “Jigsaw’s Redirect Method” will be implemented. Said method is that the moment a person searches for suspicious words they will be redirected to a playlist of curated videos that directly confront and unmask these harmful notions.
Want to know more about what is Artificial Intelligence and Machine Learning? Visit our last entry here.