Discuss Scratch

Zro716
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

FOREWORD: This is a highly radical suggestion which may require permission from big companies to actually work. Please bear in mind we're discussing if Scratch should, not will or could have a Content ID system.

More often than realized, people forget to credit something or they forget who to credit. YouTube/Google has a system called Content ID that in theory lets it process the thousands of videos that are uploaded everyday to find copyrighted content and remove it, automatically. In practice, it's heavily biased against fair use but that's another story. The gist is Google uses it in lieu of an army of moderators.

Content ID works like this: the site owner (for YouTube, their CEO) creates a database of copyrighted content received from copyright owners; when a user uploads content, their video is scanned for copyrighted material, even when compressed, slowed, sped-up, snipped, and distorted by background noise; when content ID finds a match, it notifies the copyright owner and lets them deal with the video by takedown, sharing revenue, or displaying ads. There are checks that users can take in the case that Content ID misidentified fair use (this has proven problematic way too much).

With the Scratch community growing every day, there will come a point when the ST has to make a choice: hire more moderators, or get some sort of automatic moderation in place. I reckon that at the same time, we may need to change the layout of fields on a project page - I would suggest the notes and credits be separate entities so that credits are preserved when remixing.

Although Google's Content ID system is too greedy, a similar follow-up by the ST could optimize moderation of projects without inflicting tough penalties. For example, when a user shares a project, it is compared with a database of copyrighted content (Note: let's assume Google agreed to let websites like Scratch use its database for this purpose); for every match, a piece of attribution is added to the Notes and Credits if it doesn't already contain a mention for it. So if a project has the SpongeBob SquarePants theme song, Scratch Content ID would match the same file and return the attribution string “SpongeBob SquarePants opening theme by Patrick Pinney”. If the Notes and Credits already mentioned “SpongeBob SquarePants theme” then the database check isn't necessary.

What we hope to accomplish with this is twofold: 1) making it faster/easier to credit copyright owners, and 2) better protection from copyright infringement.

First, automatic credit insertion for copyrighted content would help out younger Scratchers who aren't really aware of the importance of attribution, as well as help experienced users with crediting the right people. You can't just credit “the Internet” or “Google Images” for album covers you found, so let Scratch's Content ID match the right people and credit them instead. You might also have misnamed a sound and forgot the artist/song name, so this will help fill in that.

Second, automatic credit insertion would patch a hole in copyright issues on Scratch. If some of us remember a while back, someone brought up that there are tens of thousands of claims of copyright infringement on Scratch, which was described as a “haven for pirating”. Nearly all of the projects that violated copyrights had no mention of credit to the copyright owners, so it's reasonable to expect they got mad because of that. If Scratch checked and credited the owners of copyrighted material, I believe there would be far less takedown claims issued by copyright owners. That would be a huge sigh of relief for the ST and for innocent project creators.

I'll restate that this idea is very sweeping. It requires setting up, or borrowing Google's content ID database, and making sure attribution is given. There won't be a penalty for forgetting to credit because we'll resume the friendly reminders to do so. There won't be as much of a threat of companies suing Scratch over copyright infringement. There won't be terrible downsides if the system isn't perfect because its only purpose is to add credit. The only concrete suggestion I make is to split the Notes and Credits so that credits can be preserved when remixing, but as far as this whole post goes, I'm positing the idea of automatic credit in Scratch.
NanoRook
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

Support. Would help with people who can't bother to give credit.
jokebookservice1
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

NanoRook wrote:

Support. Would help with people who can't bother to give credit.
As per that (and for those who don't know what credit is)
Abstract-
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

No Support.
People should learn to give credit before they get sued for it at age 25.
Zro716
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

Abstract- wrote:

No Support.
People should learn to give credit before they get sued for it at age 25.
Alright, tell us how we can get people to give credit.

I mean, we have the Community Guidelines, the Terms of Use, etc.. but are these effective?

Last edited by Zro716 (May 9, 2016 23:22:18)

Iditaroid
Scratcher
500+ posts

Should Scratch have a "Content ID" of its own?

Zro716 wrote:

What we hope to accomplish with this is twofold: 1) making it faster/easier to credit copyright owners, and 2) better protection from copyright infringement.
Who is “we?”

This is an interesting idea. It's not something I would enthusiastically support, but I don't think it would be bad. An automatic system for attribution like this definitely would feel fancy!
Techno-CAT
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

NanoRook wrote:

Support. Would help with people who can't bother to give credit.
people who can't bother to give credit
_Comicfan_
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

Support. This would help with people who do not give credit. I think they should hire more moderators, though.
Tymewalk
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

No Support. This could cause a lot of problems.

I'm ignoring things like “how would it detect an image” since you said this is about if it should, not if it can.

A lot of Scratchers use memes like song clips from Vine in their projects. Some of these songs have inappropriate titles, so Scratchers simply reference it as “that meme” to avoid being flagged.

If a Scratcher uses an OK sound clip from a not-OK song, and content ID displays the song's Not OK title, that can cause problems with moderating the community. Some other Scratchers may even report the project, even though it was Content ID's fault, not the creator's.

This can be applied to ANY resource. I've seen cases when a Scratcher uses OK audio from a Not OK video, and doesn't leave a link specifically so he/she isn't posting unsafe links.
Zro716
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

Tymewalk wrote:

No Support. This could cause a lot of problems.

I'm ignoring things like “how would it detect an image” since you said this is about if it should, not if it can.

A lot of Scratchers use memes like song clips from Vine in their projects. Some of these songs have inappropriate titles, so Scratchers simply reference it as “that meme” to avoid being flagged.

If a Scratcher uses an OK sound clip from a not-OK song, and content ID displays the song's Not OK title, that can cause problems with moderating the community. Some other Scratchers may even report the project, even though it was Content ID's fault, not the creator's.

This can be applied to ANY resource. I've seen cases when a Scratcher uses OK audio from a Not OK video, and doesn't leave a link specifically so he/she isn't posting unsafe links.
Ah, that is a very good point, though I question the OK-ness of kids using parts of inappropriate media in the first place.

In that case, the content titles could be sent through the comment filter before they reach the project being shared. While being scanned for vulgarisms, if one is found, it would just throw out that piece of attribution and let the user make a ‘safe’ version instead (letting the user know that not all pieces were credited).
Tymewalk
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

Zro716 wrote:

Ah, that is a very good point, though I question the OK-ness of kids using parts of inappropriate media in the first place.
A good point, but something to be debated elsewhere.

Zro716 wrote:

In that case, the content titles could be sent through the comment filter before they reach the project being shared. While being scanned for vulgarisms, if one is found, it would just throw out that piece of attribution and let the user make a ‘safe’ version instead (letting the user know that not all pieces were credited).
True, but the bad word detector isn't perfect. It relies on other Scratchers reporting what it doesn't catch.
Plus, what if it was enabled on all projects? With 14.5 million+ projects shared, even if 100,000 projects have inappropriate resource titles, and only 1% slip through, that's still 1,000 semi-random projects that show inappropriate titles to Scratchers.
Zro716
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

Tymewalk wrote:

Zro716 wrote:

In that case, the content titles could be sent through the comment filter before they reach the project being shared. While being scanned for vulgarisms, if one is found, it would just throw out that piece of attribution and let the user make a ‘safe’ version instead (letting the user know that not all pieces were credited).
True, but the bad word detector isn't perfect. It relies on other Scratchers reporting what it doesn't catch.
Plus, what if it was enabled on all projects? With 14.5 million+ projects shared, even if 100,000 projects have inappropriate resource titles, and only 1% slip through, that's still 1,000 semi-random projects that show inappropriate titles to Scratchers.
Still, catching 99% of inappropriate instances is far more successful than not stopping any of them. The more instances it catches, the less burden of responsibility it brings on users and moderators.
Birdlegs
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

Abstract- wrote:

No Support.
People should learn to give credit before they get sued for it at age 25.
I agree with this.

Also: Isn't everything on Scratch licensed under Creative Commons, meaning it can't breach copyright? There was the Namco instance a while ago, but other than that there haven't really been issues.
Tymewalk
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

JoeyTheChicken wrote:

Abstract- wrote:

No Support.
People should learn to give credit before they get sued for it at age 25.
I agree with this.

Also: Isn't everything on Scratch licensed under Creative Commons, meaning it can't breach copyright? There was the Namco instance a while ago, but other than that there haven't really been issues.
I don't think Creative Commons is a protect-all shield against copyright infringement.
But as long as you're not selling your Scratch project with other people's work, it counts as Fair Use.
Lythium
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

JoeyTheChicken wrote:

Abstract- wrote:

No Support.
People should learn to give credit before they get sued for it at age 25.
I agree with this.

Also: Isn't everything on Scratch licensed under Creative Commons, meaning it can't breach copyright? There was the Namco instance a while ago, but other than that there haven't really been issues.
There was the whole ‘Hall of Fame’ incident…
Techno-CAT
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

Lythium wrote:

JoeyTheChicken wrote:

Abstract- wrote:

No Support.
People should learn to give credit before they get sued for it at age 25.
I agree with this.

Also: Isn't everything on Scratch licensed under Creative Commons, meaning it can't breach copyright? There was the Namco instance a while ago, but other than that there haven't really been issues.
There was the whole ‘Hall of Fame’ incident…
please elaborate
Tymewalk
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

Techno-CAT wrote:

Lythium wrote:

JoeyTheChicken wrote:

Abstract- wrote:

No Support.
People should learn to give credit before they get sued for it at age 25.
I agree with this.

Also: Isn't everything on Scratch licensed under Creative Commons, meaning it can't breach copyright? There was the Namco instance a while ago, but other than that there haven't really been issues.
There was the whole ‘Hall of Fame’ incident…
please elaborate
The Animation Hall Of Fame was a Scratch Studio a while back. However, there was also a (then planned) museum of the same name. (Animation Hall Of Fame)
The builders of the museum asked the ST to change the studio name, which at first they resisted, but then did so to avoid a legal battle.
Jonathan50
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

JoeyTheChicken wrote:

Also: Isn't everything on Scratch licensed under Creative Commons, meaning it can't breach copyright? There was the Namco instance a while ago, but other than that there haven't really been issues.
…no. When you publish a project on Scratch, you agree to license it under CC BY-SA. That doesn't mean that you're allowed to re-license nonfree works under CC BY-SA or something.
TheMonsterOfTheDeep
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

Jonathan50 wrote:

JoeyTheChicken wrote:

Also: Isn't everything on Scratch licensed under Creative Commons, meaning it can't breach copyright? There was the Namco instance a while ago, but other than that there haven't really been issues.
…no. When you publish a project on Scratch, you agree to license it under CC BY-SA. That doesn't mean that you're allowed to re-license nonfree works under CC BY-SA or something.
Yeah, as far as I know copyright law can only be circumvented with Fair Use, which is the reason behind why using copyrighted content on Scratch is normally OK.
iamunknown2
Scratcher
1000+ posts

Should Scratch have a "Content ID" of its own?

You shouldn't have used the phrase “content ID”…

Content ID is over-sensitive and has bad effects. For example, any bad company could make a bot that either 1) takes down any video that shows up on content ID, regardless of fair use or not on 2) makes money off it and takes it away from the original creators, whether that song was used as a time lapse theme that lasted 20 seconds in a 10 minute video.

Linking content ID to Scratch would open Scratch up to companies like NOMCO, where they would employ bots to scan and get projects DMCA'd (which would be way worse than a lone lawyer being mean to a single Scratcher). Sure, not as bad as suing the whole of Scratch, but still pretty bad.

I would employ another automatic moderation system like this:
  • Each user is assigned a weight, which is how much their “vote” counts. The factors of the user's weight include the time since the user has been registered, how linked they are to clouds (and how many clouds they are linked to) (more on that later), and which side (overwhelming side versus overwhelmed side) in report “votes”.
  • Reports are logged into a ledger. The system will create “clouds” of users, which are a group of users where their report histories have a lot in common.
  • The more “linked” the user is to a “cloud” + the less “vote anomalies” (more than one “loyalty” to a cloud + how big the clouds the user is “loyal” to + how much “overlap” the cloud has) that have showed up in the user's “voting record”, the less weight their vote has.
  • A voting anomaly will decrease the user's general weight based on their current weight, and the weight of the actual vote will be decreased.
  • If the ratio of reports/appeal reports are too high, the post is taken down and flagged. Vice versa as well, except that the post is granted temporary immunity instead.
  • If nothing is done to the post (immunity or taken down) for a time long enough, an option for reporting directly to human moderators will show.
  • If the users are on the weak side of a decision (when they vote for the side that loses), their weight decreases. I'd a user's post gets taken down like this, unless appealed by a moderator, the user's weight will decrease.
The advantage of the system above is that if a bunch of spambots decide to group together and spam upvotes whatever spam they had, 1) they would need to gain reputation/weight first, which would take a long time, 2) if they gradually gained weight before gradually turning spam-mode, the anomalies would

Last edited by iamunknown2 (May 12, 2016 09:26:10)

Powered by DjangoBB