Google Are Rolling Out New AI to Deal with Their Hate Speech Problem
Last month, they told their advertisers that they were developing a more sophisticated AI system to root out hateful content. In light of these new developments, they've fast-tracked it, and on Monday they released a statement saying that a lot more progress has been made. They've also gone into a bit more detail about how the system is going to work.
Essentially, Google are revamping the AI so that it can comb through YouTube's staggeringly vast library to identify inappropriate content, but they're also creating a kind of 'brand safety' reporting system so that external partners can monitor standards, something which brands have been suggesting for years.
It seems like Google could have implemented both of these measures at any time, but they never saw the issue as significant enough. With so many blue chip brands pulling their advertising from YouTube, it's safe to say they've had a bit of a wake up call, and they're response is already paying off - Johnson & Johnson have started advertising on YouTube in several major markets again.
According to the statement, in the past two weeks Google have increased their rate of flagging videos as 'non-safe' five-fold. They still maintain that, proportionately, it's a "small problem", but it seems like they're confident they can almost completely stamp it out.
It's going to take a lot more work on Google's part to win all the other companies back around, and some have suggested that brands may use the issue as a way to demand even more changes to the YouTube platform. With that in mind, some advertisers might not come back to YouTube for months after this, and some may even stay away for good. Google aren't in the habit of bending over backwards to please anyone.
What they most definitely aren't going to do is change the way YouTube functions. Using AI to weed out hate speech is one thing, but some brands would have them stop allowing uploaders determine whether or not their videos can feature ads, and changing that would hit independent producers particularly hard. YouTube's first priority is to the users, and no amount of brand bellyaching (or boycotting) is going to change that.
Callum is a film school graduate who is now making a name for himself as a journalist and content writer. His vices include flat whites and 90s hip-hop. Follow him @Songbird_Callum
Google Are Rolling Out New AI to Deal with Their Hate Speech Problem Reviewed by Callum Davies on Thursday, April 06, 2017 Rating: