LDS Freedom Forum

This is a bit wordy, so bear with me. I'll try and provide some imagesin the reples to help explain all this in replies. I notice a lot of people post YouTube content on here, and have issues with it. YouTube has been using AI learning tools to try and suss out who is a "problem" for a while. You can thank Google for that (the same company that does the search function on this site!)

Like most of these tech companies, while they do frequently ban and censor content and users openly (especially on Covid related content), they often do so in a more underhand method called shadow banning, so that the person being banned doesn't even necessarily know it has been done to them.

YouTube has had an update, in the last few days. The most obvious feature is that it shows an automated transcript of a video (with hilarious results for people with my type of accent!) However, it has come up with a new method of mangling communication and isolating comments. If you comment on a video, your comment may well be hidden from others, and then your reply will be as well.

You get notifications as to who replies to you on the app too. I've never quite worked out the logic of it, but that filters out some people too. It often notifies you about short replies or unrelated stuff, but not if you are having a regular back and forth with someone.

They also deprioritise certain content if you search for it. The content is out there, because I've seen it, but it's clear that Google as usual points you to what it wants you to see - i.e. safe, popular and vaguely in the same area. Not challenging, informative or specific. Sometimes I practically have to type in a specific title of a video, to get it, and even then...

For example, when I was looking for videos on the astronomer Fred Hoyle recently, the search was biased towards bigger YT channels, and tended towards more basic info. Even more annoying was the fact that it suggested midwit tier videos on Stephen Hawking and Carl Sagan (neither of whom I was looking for), because it obviously thinks those were a) safe and b) popular. And most people would be sucked in by that.

Likewise, if I look up content on Scottish history and culture, while I can find one or two good uploaders, it's mostly very, very basic intros written for people who don't have a clue about it.

In the first search, I looked up Kenneth MacAlpine, one of Scotland's most important kings, but not as well known to the general public as some other figueres. The results I get here are a) some AI reading out (and mispronouncing) his Wikipedia article, and b) "Scotland History Tours" doing a more general video on various Scottish kings, not just jom. I actually don't think "Scotland's History Tours" is a bad channel, and I find the presenter quite funny sometimes, but he comes up on most searches about Scottish history. He also won't teach me much I didn't know before.

In the second case, I looked up "Soisgeul Eoin" (which is the Gaelic for John's gospel). Now in the old days, if you typed in something that specific, you'd get it. You'd find a reading of the gospel. I know, because I used to do so. I realise that it has to trawl through multiple languages, but here it has recognised the language and instead of giving me John's gospel, it suggests a video about the difference between Scottish and Irish Gaelic - of no use to man or beast here. (I chose Gaelic because it is very specific - the only way this could go wrong is because "eoin" can be a man's name or also refer to "birds" - neither of these turn up.)

So this reply is basically about how content is promoted and repressed.

It's a bit harder to point out shadow banning, but I'll try. Here is a video on disability in Korea (random example). You can see there are two comments listed. Only one is visible.

You can see here, the default setting, which is "top comments" (whatever that means). Ridiculous in this instance, because no one has liked either comment. You can get this by pressing the # like symbol to the left of the X in the previous image.

(Facebook has the same trick. In replies to a post, the default setting is "most relevant" (relevant to what?) and you have to change that to view other comments. I may show that down below.)

Now we can reveal the second comment under "newest first". Denny Park has strong views about this matter, and may have been blacklisted for other comments. The words "scum", "punched" and "lowlife" probably helped get this comment hidden from the default setting.

I always look up "newest first", because many of the more interesting comments are under there. But the replies to the main comments are being censored even more extremely.

Here is a more recent example from the last few days. If someone replies to something, you can see the notification sometimes.

Here's a clip from the Blues Brothers where they drive into the Illinois Nazis. Due to the subject matter, the AI is no doubt all over this video! Someone says "do the same to black lives matter", which you can see as a highlighted reply.

I can't show you all the replies to "babiesmakingbabies" there, but I can show you this. The BLM comment is invisible in the main the main thread of replies. The laat comment visible is from three months ago.

(By the way, the Blues Brothers is not a kid friendly movie... famously it has the scene where they swear in front of the nun. A lot.)