When working on the YouTube AI Comment Moderation launch in 2016, I immediately recognized it could be a powerful tool in creating safe spaces for open conversation. I had my own assumptions about who could benefit from this tool and rallied beta testers to find proof. I saw the opportunity for Creators to be heard within their own communities and a way to reduce abuse online.
The way AI Moderation works is by holding only potentially inappropriate comments for review, and reduce the amount of time needed to monitor community comments.
As more comments are moderated, it trains the AI. This new moderation setting gets better at identifying the types of comments that creators want to review.
After getting the blessing of the Product Manager, I created and facilitated a training for a group of minority beauty vloggers at the YouTube Space in Los Angeles. The training was well received and we enabled the new beta feature on several YouTube channels.
As a minority female, I went into this training with an expectation that this tool was needed most by minority communities. Over the next month, I checked in with our beta testers and the information I received revealed an unexpected finding. There were very few comments held for review. After digging further into how the tool was being used, I couldn’t find any flaws in the setup or application. What I learned was that our minority vloggers had active communities who flagged and reported inappropriate and offensive comments quite frequently. The community was so active that they not only reduced the amount of time the vlogger spent moderating comments but they were successfully shielding their own community and content provider from abuse and trolling.
After bringing the information back to the Product Manager, I was happy that the findings were well received and not seen as a failure to drive product adoption. The reaction was to observe, reflect and apply, the behavior of a skilled Product Manager. The tool launched publicly and I still speak to it in product conversations. I still recommend it to consulting clients using YouTube.
In recent days, this has been coming up a lot for me. We need diverse teams to cover a wide range of skills and backgrounds. Together we can test as many assumptions as possible and cut down on the chance of going down a wrong path.
I’ve come to realize I was one of the diverse team mates bringing a new perspective to the product team only because I was given the room to test my assumptions. After bringing my findings back to the Product Manager, I was happy that they were well received and not seen as a failure to drive product adoption. The reaction of the skilled Product Manager was to observe, reflect and apply learnings. The tool launched publicly and I still speak to it in product conversations. I still recommend it to consulting clients using YouTube.
Image Credit: Everywoman.com, YouTube Comment Moderation