YouTube's copyright system isn't broken, the world's is (2020) [video]
youtube.com
youtube.com
YouTube assumes “guilty until proven innocent” because that’s cheaper. It’s cheaper because properly following the chain of what constitutes free speech and whether a claim is legitimate are costly questions. So it just ignores all of that and takes action because otherwise copyright trolls like the RIAA would attack them.
And that’s a big part of the problem, real world copyright laws and the DMCA are mostly used to protect large corporations. In the US it used to be 20+5 for copyright. The Founding Fathers reasoned that an artist deserved that long to profit off their successes, but that they should still innovate and create beyond that initial creation. Nowadays it’s death of the author + 75 years, or for anything produced before 1978, + 67. Walt Disney’s death was 66 years ago, with Mickey Mouse due to enter the public domain in 2024. We’re already hearing talks of Disney “losing Mickey Mouse”, so expect some intense lobbying and a bipartisan copyright bill to extend it to 100 and 87. And beyond the Death of Public Domain, we have laws such as the DMCA which are heavily abused across the entire Internet. Censoring information left and right, even when it’s perfectly legal and would be found so under the law. It’s a completely draconian idea with no balance at all.
So yes, the copyright system in general is broken, but so is YouTube. And shifting the blame does not absolve YouTube of killing off the ability to have a human review the process, especially at their profit levels. Instead, it just shows Tom Scott is an apologist for YouTube, which makes sense considering that’s where his paycheck comes from.
My impression was that the law itself was fine in this respect -- a claim could be met with a counterclaim (requiring legal commitment from the alleged offender -- signatures, identity, jurisdiction) and YouTube could then restore the media pending actual litigation. In other words, Legal Entity vs Anonymous defaults to a takedown, but Legal Entity vs Legal Entity defaults to keep up. That sounds reasonable to me, and the fact that youtube doesn't work that way is not a consequence of law but rather a corporate choice to put luxurious accommodation of large copyright holders ahead of basic decency towards content creators.
> [Copyright creep is corporate welfare]
Yep. Gross.
A counterclaim process, even if as streamlined and fully supported as possible, is still a legal operation which involves time and money that many people are not going to have, especially for small-scale and legitimate fair-use scenarios. If you copyright claim some video from Sony or IBM, they've got a legal team on retainer to get it back up. A top-tier "content creator" who's earned six or seven figures and his career depends on the video staying up might hire a lawyer and fight for it. But the guy who's using a video clip as a fair-use part of a non-commercial discussion is probably going to fold instead of going out of pocket to defend himself.
So we'd see the system falling back to "no response/takedown" for many important real-world use cases anyway.
If DMCA spam had a cost to it-- perhaps a deposit that had to be filed with the court, refundable if there is no counterclaim, or actual prosecution of bad-faith actors for perjury, that might help make it a tradeoff rather than a wholesale giveaway to the copyright industry.
Note that it only does so after a waiting period of 10-14 days after the counterclaim. There's plenty of cases where a bullshit takedown is enough to destroy the value of the material taken down.
The DMCA does not require takedown within one day. It requires “expeditious” removal. But courts have held that response times ranging from 5 to 14 days are expeditious, and in a recent case against YouTube (Business Casual v YouTube), a 23-day delay was excused because the court applied the standard for purposes of volitional conduct and direct infringement, sidestepping the DMCA entirely.
Some quotes:
The plaintiff “has not pointed to any authority to support the proposition that YouTube was under a legal obligation to conduct its investigation into Business Casual’s complaint on a more compressed timeline”
“Business Causal has failed to claim plausibly that any purported delay by YouTube constituted active, volitional conduct that caused TV-Novosti’s alleged infringement”
That case was significant for a bunch of other reasons, too. See discussion at https://blog.ericgoldman.org/archives/2022/03/youtube-defeat...
That is a very interesting case! The part about the DMCA not applying to most Youtube takedowns (because the legit uploader has given Youtube a license) is particularly surprising, despite being a straightforward implication of the law.
The licensing aspect is the surprising part I was referencing, but I understand it a bit differently. It’s not that the DMCA doesn’t apply. Rather, it’s that when you upload your content to YouTube, YouTube’s terms of service grant YouTube a broad license to use the content you uploaded. Consequently, YouTube can’t be liable for directly infringing any copyrights associated with that content. Since you licensed it to them. The DMCA still applies.
The court states this in a bit of a roundabout way (compared to my blunt summary):
> The DMCA safe harbors provide potential defenses against copyright infringement claims where, but for the safe harbors, the plaintiff has a meritorious cause of action against the defendant for copyright infringement. But alleged failures to satisfy the conditions of a DMCA safe harbor provision cannot constitute a cause of action without a viable underlying claim for copyright infringement.
Since youtube is incorporated in the US all youtube users worldwide would have to submit to US jurisdiction anyways? (On top of whatever country they're in)
Whether that falls under US jurisdiction, another country if they're outside, both, etc... for the case of youtube hasn't been settled yet. Though it seems very likely to be trending towards that direction.
We're specifically talking about legal jurisdiction. An instance of copyright infringement depends entirely on jurisdiction, because copyright is nowhere near any sort of natural law.
I agree that practically the US government will attempt to invent some rationale to bring every action in the world under US jurisdiction (having looked at a Federal Reserve Note once in your life, wearing a shirt that is colored red, white, or blue, etc). But assuming other countries might actually stop acquiescing to this type of nonsense some day, then an individual having explicitly assented to it is likely to be a strong point in favor of continuing to asserting worldwide US jurisdiction.
This is false for the many countries that have signed on to various binding international IP treaties, such as the various WIPO treaties which the US has also ratified. You should read up on it.
A copyright issue involving an online service such as Youtube in one WIPO ratifying country would near-automatically translate into an issue across many jurisdictions.
China also has ratified the 1999 convention as of last summer, so the area of the Earth where someone can ignore these treaties is shrinking rapidly.
Haha thanks for making me feel old. I suspect many of us who remember life before DMCA don't consider it halfway reasonable.
The Copyright Act of 1831 was an extension of the original copyright term from 14 years to 28 years, with the same option to renew once for another 14 years, or 42 years. It takes until 1976 for copyright to start to extend into incredibly long lengths of time.
The last update to Copyright law 24 years ago was also post DMCA being established. Seems hardly a coincidence that both go hand in hand as being unwieldy to non corporate entities.
Good luck avoiding Disney's trademark lawsuits.
If a creator wants to ignore the YouTube's copyright strike system entirely, and YouTube will get rid of the strike, but that also means that you're accepting the legal liability instead of YouTube.
Fran Blanche got a strike for a film with the sound of blowing wind. She's gotten many other strikes from public domain content.
Sony claiming the sound of wind: https://youtu.be/3TFESXDVdhM
Troll claiming ownership of PD NASA films: https://youtu.be/1qCM9L_FaFU Discussion: https://news.ycombinator.com/item?id=32159830
From google:
> Also, Content ID claims don't result in a strike.
Again, if you want to get rid of a strike, you just have to submit a counternotification. If it's obviously bullshit, then you'll get your video reinstated and you're probably not going to get sued. It sucks that you have to deal with this, but the harsh truth is that you just have to deal with stuff like this as a business owner.
I'd wager that the YouTube EULA makes it clear that fraudulent YouTube takedowns do not create liability, and therefore protect the trolls.
Take downs ought to required to be attached to a substantial bond per creator not per work that can be recovered in country without substantial litigation. Legit creators could act through an intermediary while paying only a small fraction of the bond under the premise that they would rarely lose the bond and could recover from someone who has a legal identity attached.
Trolls would virtually cease to exist as it would be expensive and futile. Post one fake notice lose your money exit.
Legit companies that presently haphazardly take down programmatically would adjust their strategy after they lost the nth bond.
To quote them:
> "[YouTube] goes far beyond its legal obligations in assisting content owners to protect their works."
But media companies didn't like far beyond legal obligations, they wanted more.
And existentially, maybe YouTube was right to do that. Today, YouTube is clearly profitable and it's more culturally relevant than television in many places. But 10 years ago, it was edgy and new, scrappy, it didn't have as much mainstream appeal, and even if they ultimately won all of those lawsuits eventually, it would've left them with a ridiculous amount of legal spending and a bunch of possibly permanently severed relationships with the media companies that had all of the money. I don't think Google wanted to just delete channels of everybody. I think they did want to delete channels of people who were just splitting Family Guy episodes into 15 minute chunks and uploading that, but definitely there's a reason why copyright claims on YouTube today have more cardinality (depending on the kind of claim, for example, it can result in ad revenue going to the claimant, geoblocking, etc. Which also can be bad outcomes, but is a better outcome than just demolishing people's channels.)
Google is a lot more on the evil end of things today than they ever were, but back when all of this went down in 2010-2014 it wasn't so bad. I think YouTube's compromise reflects more than just problems with Google. I think they developed a copyright system that would prevent them from being in constant legal turmoil indefinitely.
The only official number from Google is $15 billion in revenue or 10% of their total revenue. But that’s not much once you consider hosting and traffic costs.
https://www.theverge.com/2020/2/3/21121207/youtube-google-al...
YouTube would be a horrible standalone business.
The way the DMCA was supposed to work, the claimant submits a notice and the site takes it down, the user submits a counternotice and the site puts it back up, and now if the claimant still wants to proceed, the site has nothing to do with it anymore and they go to court against the user.
But the users don't have any money, and there are tons of them, so instead they went to court against the site. Which is the exact thing the DMCA was supposed to prevent.
Part of the reason for this is that pre-Google YouTube did some shady things that gave them an excuse. But then we have the other problem with the law. If you screw up, the penalties are unreasonable, which allows litigious plaintiffs to extract unreasonable concessions. And the same is true even if you didn't screw up, because the alternative is to spend a crazy amount of money on lawyers because the stakes are high.
When the The Copyright and Related Rights Regulations 2003 (https://www.legislation.gov.uk/uksi/2003/2498/regulation/8) is repealed at the end of 2023, because it is Retained EU Law pursuant to an E.U. Directive (https://www.legislation.gov.uk/eudr/2001/29/article/5) made under the authority of the European Communities Act 1972, and making temporary copies in HTTP proxies and WWW browser caches becomes illegal again in the U.K. (which was one fault in a 1988 Act that the 2003 regulations were fixing), I wonder whether Tom Scott will notice.
If you look at s1(6) of the Retained EU Law Act [1]:
> The revocation of an instrument by subsection (1) does not affect an amendment made by the instrument to any other enactment.
If you read the Explanatory Notes to the Commons Bill, it states this at paragraph 71 on page 11 [3]:
> Subsection (3) clarifies that any amendments made by EU-derived subordinate legislation and RDEUL are not within scope of subsection (1). For example, where an instrument made under section 2(2) ECA 19721 is revoked by the sunset, if that instrument amended primary legislation, the amendment in the primary legislation will be unaffected.
(Subsection 1(3) is now subsection 1(6) due to amendments, but the text of the subsection is unchanged.)
The Copyright and Related Rights Regulations 2003 amends the CDPA - and the particular clause you refer to has been inserted in to the CDPA as s28A. Were the regulation to cease to apply due to s1 of the Retained EU Law Act, it wouldn't matter because all it does is amend CDPA so falls within the amending exception.
As the Bill passed through Parliament, it was amended to require the Government to fully explicate all the legislation that would be affected and include it in a Schedule to the Bill. If you check the schedules, the CRR isn't on there.
The government have published a commented list of this schedule as an OpenDocument spreadsheet which gives a little more explanation of why they're repealing each particular one. [3] If you download it and CTRL-F for copyright, the only entry is the Artist's Resale Right (Amendment) Regulations 2009 which it lists as superseded by the 2011 regulations but not formally repealed.
The Retained EU Law Act is likely to have some important effects, and there are some important legal provisions like the Working TIme Directive and various bits of environmental law that will be going through the process in s1, but on copyright, it does nothing. Certainly not repealing s28A of CDPA.
[1] https://www.legislation.gov.uk/ukpga/2023/28/pdfs/ukpga_2023... (it's not been HTML-ified yet...)
[2] https://publications.parliament.uk/pa/bills/cbill/58-03/0156...
[3] https://www.gov.uk/government/publications/schedule-of-retai...
It's a design choice by Youtube.
But maybe it was just a realization of what was more profitable? R&D becomes exponentially more expensive the more advanced search you try to provide, but any random weird company will be interested in user data for marketing.
Just a little anecdote: I have subscribed to a lot of newsletters about some FOSS projects I like. I use to never read those letters and now I have reached the limit. Now I can not delete those newsletters without deleting everything because I can not mark more than few tens of those letters at once while having few millions of them.
if you labeled them or create a search that is specific to them ( perhaps from:mailing-list@address ):
there is a checkbox above the email listing. clicking it selects everything on the current page.
instead of clicking that, click the arrow next to it. a variety of selection options appears.
select "all"
a banner then appears that says you've selected everything on the current screen, and offesr to select everything that matches it in your account
click the offer to select all matching items
you can then issue a mass move or mass label or mass delete or whatever
I haven't tested this (I'm not purging my own mailing lists to try it :) ) and a selection mistake would be costly, but it appears to be correct.
I guess people who are complaining should learn how to do the skilled trades - electricians, plumbers, HVAC, car repair.
Microsoft got so bad I had to switch operating systems (no real love lost, but I really did enjoy XP, 7 and even Vista for nostalgia reasons - first desktop PC), and Meta... do I really even need to say anything?
They most certainly are not...
If anything, Google is trying it's damned hardest to pull off a "Microsoft v2" party.
Everything from intentionally harming competitors, defrauding their ad customers, ruthless tracking and selling information about their users, hoovering up the competition and then killing it (or at best, letting it rot in some corner). And oh - by the way! Google also happens to own the dominant web browser and is merrily InternetExplorerifying it more and more every day.
Frankly - the resemblance is so strong to mid 2000s MS that I'm staggered.
This can also be said for VALVe since their dev team also does customer service (at their scale).
(I wish this whole video was done by Jay, he is infinitely more entertaining and less patronizing sounding)
(The post should have 2020 in the title)
(but it is relevant today bc Tom's remarks about casual theft of photographers and other creators hard work directly apply to AI training. ClosedAI is basically the Giphy in his video)
Does it in fact address all the issues that Scott raises in his video?
<https://web.archive.org/web/20230323113746/https://old.reddi...>
I've noticed many people:
- Fail to acknowledge a problem at all.
- Come up with the wrong diagnosis (mistaking symptoms for syndrome, or falsely attributing root cause). Also, as you have, identifying a part but not the significant whole of the problem.
- Leaping straight in with a preferred method, a/k/a the "if you have a hammer..." syndrome (everything is a nail), or "garbage can theory" (you'll apply the set of tools in your toolbox to every problem, regardless of whether or not they fit).
The other stages, successful execution, coordination (which often significantly involved preventing others from blocking effective measures), and assessment, also appear, but typically are mooted by issues further up the chain, or apply to grander-scale problems.
Watch the video if you have the chance. It's good and thoughtful.
And it really doesn't involve crowdsourced work to much degree.
Edit: Swapped Wayback link for origin which is ... on RedditStrike.
The title simply suggests a completely different story and most people don't have the time to have their prejudices refuted.
Up to now, no one has given a concise summary of this video in this submission.
I'm well aware that video and audio pose some challenges to ready consumption or perusal. That said, Tom Scott is among the more effective and informative YouTube creators out there, and I'd argue that this is among his best videos as well.
(Another that comes instantly to mind is his Royal Society lecture on truth and epistemology on the Internet.)
I'm also painfully aware that YouTube's policy shifts of the past day or so have made accessing and viewing content there even more obnoxious than usual.
Still, argument from ignorance tends to have weak foundations.
(That subreddit is my own personal blog, and I'd set it private myself. Those responsible are being harshly punished.)