Alphabet Google YouTube: We Are Doing Darned Good Work

April 7, 2021

I read a peculiar item of information about the mom-and-pop outfit Alphabet Google YouTube. You may have a different reaction to the allegedly accurate data. Just navigate to “YouTube Claims It’s Getting Better at Enforcing Its Own Moderation Rules.” The “real news” story reports:

In the final months of 2020, up to 18 out of every 10,000 views on YouTube were on videos that violate the company’s policies and should have been removed before anyone watched them. That’s down from 72 out of every 10,000 views in the fourth quarter of 2017, when YouTube started tracking the figure.

Apparently the mom-and-pop outfit calculates a “violative view rate.” This is a metric possible only if a free video service accepts, indexes, and makes available “videos that contain graphic violence, scams, or hate speech.”

The system, the write up reports that :

YouTube’s team uses the figure internally to understand how well they’re doing at keeping users safe from troubling content. If it’s going up, YouTube can try to figure out what types of videos are slipping through and prioritize developing its machine learning to catch them.

A few questions spring to mind:

  • What specifically is “violative” content. An interview I conducted with a former CIA operative was removed a year after the interview appeared as a segment in my 10 to 15 minute twice monthly video news program. An interview with a retired spy was deemed violative. I hope YouTube learned something from this take down. I remain puzzled.
  • How does content depicting graphic violence, scams, and hate speech get on the YouTube system? After I upload a video, a message appears to tell me if the video is okay or not okay. I think Google’s system is getting better from the mom-and-pop outfit’s point of view. From other points of view? I am not sure.
  • Why trust metrics generated within the Alphabet Google YouTube outfit? By definition, the data collection methods, the sample, and the techniques used to identify what’s important are not revealed. FAANG-type outfits are not exactly the gold standard in ethical behavior for some people. I, of course, believe everything I read online like transcripts of senior executives’ remarks to Congressional committees?
  • Why release these data now? What’s the point? Apple is tossing cores at Facebook. Alphabet Google YouTube is reminding some that Microsoft’s security is interesting. Amazon wants to pay tax. Maybe these actions and the violative metric are PR.

The write up contains charts. Low contrast colors show just how much better Alphabet Google YouTube is getting in the violative content game. I love the violative view rate phrase. Delicious.

Stephen E Arnold, April 7, 2021


Comments are closed.

  • Archives

  • Recent Posts

  • Meta