The Youtube Science Spam Problem

[jedishrfu]

Kyle Hill has done a rough analysis of Youtube Science channel content and says that there is a spam problem at Youtube where folks are creating content at an astonishing rate of poor quality science videos for monetary gain.

 

[jack action]

I've seen this type of content and it is easily recognizable and you quickly learn how to ignore it ... just like we all do with ads.

I see this as a problem for Youtube. If there is too much of it, there will be no point in using the website.

But if that is what people want (and pay for), who am I to tell them what to watch?

 

[Vanadium 50]

Folks are creating content at an astonishing rate of poor quality science videos for monetary gain.

Unlike, for example, journal publications?

 

[Grelbr42]

There is certainly a significant amount of low quality on YouTube. But, I don't know what could reasonably be done about it. The AI isn't quite up to the task as yet. Having humans do some kind of check would be excessively time consuming and require people that knew the difference.

And there are questions about whether it is reasonable to even try it. YouTube, and other web sites where regular citizens can just post stuff, are supposed to be part of "the town square." That is, it is supposed to be a lot like private conversations that might happen in people's ordinary face-to-face life. While it might be desirable for such conversations to be of high quality, what sort of authority would it be that tried to police them or enforce some sort of quality over them? Consider whether you are generally comfortable with there being such an authority. Project how it might operate and what it might use as enforcement. Consider the potential for abuse by both "well meaning" idiots and people with ill intent. These thoughts should give you some reason for concern.

I can see some potential value in a clearing house sort of thing that lists actual scams. Something like Snopes is supposed to be and sometimes is.

 

[BillTre]

It would be like playing wac-a-mole to try to ding all the bad videos, especially if they are being generated so fast.
Unlikely to work and would be expensive.

Instead, a system that positively labelled the good videos (which should have an easy search method on youtube) would let people pull the good ones out of the heap. You can then ignore the bad ones.
It would have to involve more than self-identifying as good (meaning some outside group, which would have an expense), but the costs could be covered with earned youtube money.

 

[Grelbr42]

Hmmm... The vids are being created by a basically automated process. I need an AI to preview them and only offer to me the ones that have some useful content.

 

[phinds]

there is a spam problem at Youtube where folks are creating content at an astonishing rate of poor quality science videos for monetary gain.

I'm shocked. Shocked I say !

 

[pinball1970]

YouTube is great for music.
Some other good stuff but plenty of rubbish too.

As a comparison one of the smaller libraries in the area has a big section on, self healing, eating for healing, quantum healing and star sign healing.
Just walk past.
Same for those sites.

 

[DaveC426913]

Everything Grelbr42 said!

 

[DaveC426913]

Instead, a system that positively labelled the good videos (which should have an easy search method on youtube) would let people pull the good ones out of the heap. You can then ignore the bad ones.
It would have to involve more than self-identifying as good (meaning some outside group, which would have an expense), but the costs could be covered with earned youtube money.

How would this work though? Who decides? I guess you're suggesting a Youtube equivalent of Consumer's Report magazine? (And that still doesn't stop the million low-quality "curator" copy cats that will pop up to swamp it.)

 

[BillTre]

How would this work though? Who decides? I guess you're suggesting a Youtube equivalent of Consumer's Report magazine? (And that still doesn't stop the million low-quality "curator" copy cats that will pop up to swamp it.)

I would not want a bunch of youtube employees doing it. Too much possible conflict of interest.
Small outside groups could review things within their area of expertise. There would be many possible groups of reviewers which could possibly form a larger confederation of reviewers.
The reviews would be at the request of the video makers. Lousy videos would probably not have a review requested (reducing workload).

If I were organizing everything, I would have vetting done by knowledgeable groups of people. Something like a bunch of volunteers from Science groups like the American Genetics Society or AAAS.
Eventually, expenses could be paid by the youtube funds to the videos deemed good.
Getting good volunteers would seem to be the big hurdle to me.

 

[DaveC426913]

Getting good volunteers would seem to be the big hurdle to me.

I think the problem here is that it will always be subject to corruption. Pretty quickly, organizations with vested financial interest will sponsor their own "volunteers", and we'll be right back to a vetting model driven by money. I don't know how to avoid that. (Well, Wikipedia has done it, I guess.)

 

[pinball1970]

I think the problem here is that it will always be subject to corruption. Pretty quickly, organizations with vested financial interest will sponsor their own "volunteers", and we'll be right back to a vetting model driven by money. I don't know how to avoid that. (Well, Wikipedia has done it, I guess.)

Wikipedia is not that bad though? For established science?
Trickier on other stuff but what should we expect? It's trickier!

 

[pinball1970]

I think the problem here is that it will always be subject to corruption. Pretty quickly, organizations with vested financial interest will sponsor their own "volunteers", and we'll be right back to a vetting model driven by money. I don't know how to avoid that. (Well, Wikipedia has done it, I guess.)

These guys could come up with something?

Or a tech version.

https://www.cochrane.org/about-us

 

[Rive]

I would not want a bunch of youtube employees doing it. Too much possible conflict of interest.

Sure. Stupidity, debunking and debating brings far more AD views than some dry but correct content.

 

[BillTre]

The idea would not be to tell everyone this is the only thing for you.
Rather it would be if you want competant science, search for this term of certification.
The choice is the viewer's.
If they want just:"Stupidity, debunking and debating" then don't search for that do something else.

 

[Rive]

I do understand what you are after. I'm just warning you that it's an uphill battle. Youtube/google (or any other company of this business) has serious money from all that ... stuff. Your nice, clean scientific certificated videos will still be littered by really colorful recommendations

 

[phinds]

I'm just warning you that it's an uphill battle.

Oh, I don't think it would be that easy. An uphill battle CAN be won. The boat has already sailed on making the Internet a rational place.

 

[Frabjous]

Given the amount of debate that occurs at PF about what compromises are acceptable for science popularizations, I do not have high hopes for such an effort.

 

[russ_watters]

Unlike, for example, journal publications?

The difference is that reporters want a steady paycheck and youtube content creators just hope for one. AI will fix that.

 

[jedishrfu]

They don’t need to restrict video content only provide a means for serious viewers to get peer reviewed video content. One way it could work is by creating a badge for good content providers awarded by those who already have the badge and feel the content is STEM worthy.

Perhaps they could create a separate website that references the good videos.

Perhaps PF could initiate such a scheme using our membership to review and nominate good videos and good video content providers.

Hey @Greg Bernhardt

 

[Tom.G]

Perhaps PF could initiate such a scheme using our membership to review...

Sounds like @jedishrfu just volunteered!

 

[James Demers]

I routinely block channels that offer garbage content. There is a bit of a whack-a-mole problem, but it does keep things my "recommended" videos column relatively sane.

 

[Grelbr42]

Wikipedia is not that bad though? For established science?
Trickier on other stuff but what should we expect? It's trickier!

I would say, if it is a situation that matters, do not trust anything on WP. Never use it as a primary source. Sometimes you can use it to "get up the learning curve." But basically, keep in mind that it is created by a bunch of "wander-by" volunteers. I trust it about as much as somebody I don't know that I meet in the hallway at the office.

Anything on WP with any controversy at all, and I mean the tiniest bit, should be taken as highly suspect. Anything with any political involvement is likely to be pretty much horrible.

Anything technically involved should be examined exceedingly carefully. Even the pictures can be misleading. I once got into some professional trouble when I trusted the diagram on WP for the decay of Cobalt 60. Go for the primary technical sources.

 

[DaveC426913]

Sorry, I wasn't suggesting the content on Wikipedia is, itself, reliable scientific content, I was simply suggesting that their model for curating content seems to be doing a good job of avoiding the pitfall of sponsoredness and other forms of private manipulation.

eg. An article on dairy milk production might have some occasional specific errors in it, but as far as it seems to go it's not like it's been taken over by - and is thus a shill for - the Dairy Farmers of America organization. (Not sure I can defend that though - or how I'd even measure it.)

 

[Vanadium 50]

an article on dairy milk production

You should see the "Holsteins are better than Guernseys" fights.

But that's really the problem - Wikipedia's model is that the experts will drown out the crackpots, no matter how loud the crackpots are. But a few impassioned nutjobs with nothing else to do can make a lot of noise indeed.

 

[Rive]

Anything on WP with any controversy at all, and I mean the tiniest bit, should be taken as highly suspect.

Actually, I would recommend the same for all content without any controversy

... just to put it in context: based on local experience bored people can and will comment on almost anything.
Even things like 1+1=2 are not exceptions.
... so if something is still without any controversy, extension, comment or such, then it's very likely that nobody ever checked it through => highly suspicious