Web giant Google claims that a new AI system has increased the number of YouTube videos identified as containing hate speech by a factor of five.
The company has seen a number of major advertisers withdraw their YouTube campaigns in recent weeks after growing discomfort of the association between brands and racist content, much of which seems to be existing unmonitored on the video platform.
Now Bloomberg reports that Google has employed new systems that have significantly improved its hit rate for catching offending content. These new filters are designed to disable ads on videos including "dangerous and derogatory content" as well as content that promotes negative stereotypes about targeted groups or denies "sensitive historical events" such as the Holocaust.
The company even claims that advertising partner Johnson & Johnson has reversed its decision to pull its ads on the back of its success.
"We switched to a completely new generation of our latest and greatest machine-learning models," chief business officer Philipp Schindler said. "We had not deployed it to this problem, because it was a tiny, tiny problem. We have limited resources.
"Although it has historically it has been a very small, small problem. We can make it an even smaller, smaller, smaller problem."
Google also claims that toxic content has also appeared alongside around one one-thousandth of a percent of YouTube’s total ads.
Note, however, that Google admits that alongside its new AI it has also employed "a lot more people" for content checking, again asking questions of the new system’s long-term viability.
It was reported last week that a growing list of major global brands were withdrawing their advertising spend on YouTube due to the fear of being associated with hate speech.
The Google-owned company has been rocked in recent weeks by the increased awareness of the racist hate speech that has become alarmingly common among some of its top stars. Gaming has been hard hit by controversies including the likes of PewDiePie and JonTron, although of course the problem of games pundits espousing extreme right wing views has been around for some time.
Furthermore, the yanking of YouTube advertising is having a wider effect on Google’s entire ad network, with many companies defaulting back to just targeted search advertising as they no longer trust Google’s ad algorithms.