1 Unsafe video every second? That’s just the beginning!
Brand Safety (ensuring ads do not appear near harmful or inappropriate content) has been everywhere this week.
With advertisers and users demanding greater transparency, Facebook decided to externalize an internal document containing standards for what content is allowed and what is not allowed on Facebook. The document explains, for example, that nudity is prohibited, in both photographs and illustrations, unless it is posted for educational, humorous or satirical purposes. Other exceptions include nudity for art or political protest, as well as active breastfeeding.
Facebook also confirmed the existence of a massive manual content review operation – “Right now, we have more than 7,500 content reviewers, over 40% more than the number at this time last year.” This number is expected to swell to 20,000 soon. And Facebook even announced a new process to appeal content that has been removed from the platform.
On the Google front, CNN exposed, once again, major Brand Safety issues on YouTube. Hundreds of brands unwillingly ran ads on YouTube channels with content featuring problematic content – in this case featuring white nationalists, Nazis, pedophiles, and North Korean propaganda. In an example of perhaps the least suitable content-ad alignment imaginable, ads for the Jewish National Fund and Friends of the Zion Museum ran against an anti-semitic clip alleging Jewish domination.
Just days later, Google released its Transparency Report, an interesting document that details and breaks down data about 8.3 million inappropriate videos removed from the platform in Q4 of 2017. If you do the math, that’s more than one video removed per second!
According to the report, more than 6.6 million clips were identified by Google’s AI tools (which the company admits are limited), 1.1 million by Google’s (human) moderators, and 400,000 were flagged by viewers. While three-quarters of the videos were removed immediately, 1.6 million clips were live at some point and viewed at least once before being removed from YouTube.
And with Brand Safety clearly becoming advertisers’ #1 concern, it is no surprise that Oracle announced on Tuesday that it is acquiring Grapeshot for an undisclosed sum to help protect its customers from inappropriate content and to contextually target textual content. With that, Oracle gains a leg up on other marketing & data giants.
I believe this week is a microcosm of what is to come – as long as Brand Safety continues to be top of mind, we will see more Brand Safety missteps and outcries, growing pressure on platforms to provide greater transparency, better Brand Safety technologies developed in response, and this will lead to additional M&A activity by those trying to protect brands from unsafe content. Buckle up!