Here is the deal. Twitter probably uses machine learning systems to identify videos of child sexual exploitation.
In 2012 Google showed how computers can accurately identify certain types of videos. Google showed how computers could identify cat videos. Since 2012, this technology expanded. All of silicon valley uses machine learning systems for content filtering.
In a nutshell, these systems identify features and attributes in a video, compare it to known videos and make a computer decision of the type of video it is.
Twitter’s algorithm to find ‘child sexual exploitation’ would probably identify the age of the people in the video as one step. Look at how the two people interact as another. Process other deep learning techniques and slap a label.
That brings us to twitter flagging C-SPAN clips of Joe Biden interacting with with young women.
User @koenigjake said he tried to post “video of evidence on how Biden interacts with woman and little girls” but Twitter flagged him “for trying to post images of child molestation.” The fact he wasn’t able to respond suggests the automated filters identified the type of video ON UPLOAD. Also, if it were a manual review, it’s unlikely Twitter would label the video ‘child sexual exploitations.’
Twitter likely scanned the video, applied machine learning techniques and slapped the ‘child sexual exploitation label’ preventing this user from posting the video. This would mean an unbiased observer (advanced computer systems) concluded what Biden is doing in this videos is not normal behavior.
In @koenigjake’s follow-up post, he shows a C-SPAN video screenshot.
Here is the video. You decide.
We won’t get into the debate. But we’ll post the video so you can decide for yourself. The videos are all public videos on C-SPAN.