Also the DMCA doesn't require Google to takedown sites based on obviously fraudulent DMCA notices.

It's just that Google would lose its liability shield if the claim turned out to be real.

i thought that was the main flaw with the DMCA - it actually does require you to obey all properly formatted and submitted DMCA notices, regardless of how obviously fraudulent they are, until a court decides they're not fraudulent.

DMCA doesn't require you to do anything, because it only shields you against liability you would otherwise have to the filer (or the owner whose agent they are), and if the notice is obviously not from thr owner or agent, or obviously incorrect about the content being infringing (or even existing), then there is no source of liability.

Yes, it's cheaper not to check, but its not required by the DMCA.

Honest question, Does the DMCA allow hosts to require proof before action?

If their liability isn't contingent on being provided valid evidence beyond a simple statement, then hosts don't have much of an ability to perform an assessment.

> Honest question, Does the DMCA allow hosts to require proof before action?

The DMCA (well, the Safe Harbor provision under discussion, the DMCA has lots of other provisions that are irrelevant to the discussion) doesn't require action in the first place, so it necessasrily allows anything before action.

> If their liability isn't contingent on being provided valid evidence beyond a simple statement, then hosts don't have much of an ability to perform an assessment.

The DMCA doesn't create liability in the first place.

What stops the Google's of the world from simply ignoring every dmca takedown request? Alternatively what do you think prevents them from requiring rigorous proof in a machine readable format before honoring a request? What do you think the incentives at play are?

> What stops the Google's of the world from simply ignoring every dmca takedown request?

The same thing that leads them to make deals with big money copyright-based industries for copyright systems that go far beyond what the DMCA safe harbor requires for situations it covers: the fact that they are deeply and actually aware of the copyright violation they facilitate (putting them outside of the DMCA safe harbor to start with), and rely almost entirely on the fact that that is hard and expensive to prove (which the DMCA negates, but which forces them to go beyond the DMCA for industries for which the cost and expense of proving actual knowledge would otherwise be worthwhile.)

The lawsuit with Viacom in the mid 2000s that could have basically killed the entire platform over copyright complaints with big studios. ContentID was a direct result of that lawsuit with the explicit goal of making sure google would never be held responsible for copyright violations.

The vast majority of "Copyright claims" on youtube are not DMCA takedowns.

Google doesn't require proof or ask for evidence or anything because they don't give a fuck about the rights of a creator. The entire goal of youtube's copyright programs are about the continuation of youtube as a profitable entity. They do that by purposely making it easy for big rightsholders to get what they want, and in the process have made it easy for strangers to bogus claim things and possibly get advertising revenue for a bit unless the creator has enough clout and access to youtube support reps to get it fixed.

Don't big rights holders also care about having their content taken down?

Google's primary customers are rights-holders: particularly giant media corporations and very-popular youtubers.

Shouldn't giant media corporations also care about illegitimate take down notices against them?

All of their content is on a totally isolated platform like Netflix or YouTube Red.

The law binds the out group to profit the in group.

> What stops the Google's of the world from simply ignoring every dmca takedown request?

They become liable for copyright infringement if the work in question is actually in violation of copyright.

You’d want to be very sure that a work you were hosting wasn’t in violation of copyright before outright ignoring a DCMA request.

Yeah, and what is funny, is that it totally misses content on smaller artists while incorrectly IP on older, open-domain, classical music performed by smaller artists (think high-school bands performing Beethoven).

Google, nor any other corp, should be a Judge on a legal matter. That's why a Judge should be involved.

The entire idea of of the DCMA is just massively flawed and biased towards a world where only big publishers and big criminal enterprises exist.

There are 600 million DCMA notices to Google along every year. Let's be insanely Conservative and assume there are a billion across all platforms.

Let's say a judge can review a notice every 5 minutes. 12 an hour, say 100 in a day. Say 20 000 per year. So 50 per million, 50 000 judges per billion requests.

Let's say a judge costs 100k per year. With say another 100k in costs. That's a price tag of 10 billion per year. I'm guessing the real number is likely well over double that.

So your plan to include humans, never mind actual judges who went to law school and so on, is, well a non-starter.

This is an argument to re-write the DMCA to prevent frivolous claims, and thus reduce the total number of claims. It's not an argument to throw up our hands and pretend that it's unreasonable to expect human involvement.

I don't think anyone is advocating for the status quo, but that doesn't mean human involvement is required either at an initial stage.

One proposal is that you should need full legal identification to file a takedown claim, and be on the hook for damages and penalties if is malicious. This would solve 99.999 of the problem.

> One proposal is that you should need full legal identification to file a takedown claim

How exactly? Require a government issued ID? What if it's fake? How would you validate that the ID is authentic? Even if it is authentic, how do you ensure that person has proper rights to issue the takedown?

> and be on the hook for damages and penalties if is malicious.

How? They could either not be who they say they are and/or not located in the United States. Like in this exact case, the two defendants are located in Vietnam. There's zero chance they're going to show up, so it'll result in a default judgement that will never be collected on.

The best suggestion I've seen so far is to require an escrow deposit on takedowns that is forfeit on fraudulent/malicious claims. However, this then raises the issue of who would determine that. Also, this deposit could potentially tie up a lot of money of legitimate claimants, becoming a financial burden for them and preventing them from issue further claims (which further adds to that burden).

> and be on the hook for damages and penalties if is malicious.

Honestly, I think both conditions might be met with an already vetted credit card number, which obviously involves an associated identification of a person or company.

When signing up for cloud service providers, I'm always terrified that I'll leave something on and incur a massive bill. Basically, every cloud service move I treat as if I'm walking on eggshells. I'd assume those issuing a DMCA would end up the same way.

>How exactly? Require a government issued ID? What if it's fake? How would you validate that the ID is authentic? Even if it is authentic, how do you ensure that person has proper righgts

There is an infinite list depending on how strict you want to be. They could require government ID. They could require a notary. They could require a court order.

There will always be a balance between ease of takedown for legitimate claimants vs fighting false claims.

> How? They could either not be who they say they are and/or not located in the United States.

Require either US Identity papers or a corporate identity that is registered in the US in order to file a DMCA (or <country here> id/corpid for filing a <country specific> copyright claim)

Furthermore, tweaking the reporting times for DMCA would help: 1. Claim made, soft-takedown immediately (delist but don't remove) 2. Proceed to hard-takedown after 24 hrs if no counterclaim is made. 3. Counterclaim made, reinstate and inform original claimant. 4. Original Claimant can then either sue or obtain a court-ordered injunction. 5. Optionally Claimant can pay a nominal fee for "human decision-making" by Google or a mutually agreeable arbitrator. 6. Respondent has 14 days to file their own nominal fee to move it to arbitration or can proceed to countersue.

I like this idea. It adds enough friction to scare away most scammers.

This would pretty much entirely remove the ability for regular people outside the US to use the DMCA while corporations to abuse the system. Corporations don't really need that help and can get a court to approve an injunction instead - if they don't already have a business relationship with the host they can levarage to get the content taken down.

No, requiring any party to be damaged based on accusations alone is absurd. Copyright infringement is not a life or death matter. It can wait for a judge to at least look at the matter. And if that is too expensive for society then maybe reconsider copyright instead of externalizing the cost required to maintain the fiction that information can be scarce.

It needs to be easier to claim damages from inaccurate takedowns. Not only do you need to prove that the claimant knew his takedown was false, you also need to demonstrate "tortuous business interference".

A simple, fixed price minimum damage for inaccurate claims would solve this. No more robo-claims because a forum user mentioned keywords which a poorly written scraper matches to their (brokered) client's intellectual property.

Flood the system as a form of protest.

You mean you actually have to pay a significant number of humans to run a giant monolithic content platform responsibly?!

But our profits!

Petty technicality, but most jurisdictions in the USA don't require judges to have gone to law school, passed the bar or even know how to read.

Yup, that's about right for enforcing the current system of copyright, at the current level of enforcement, but with better than piss poor accuracy.

Frankly, we're over-enforcing copyrights, which is why this looks ridiculous. It doesn't make economic sense to try a sane system for most of these complaints because we'd be spending thousands of dollars on court costs to take down a webpage that gets 10 views per day.

There should be a nominal fee to file a DMCA complaint just to prevent wasting everyone's time with enforcement actions against pages that no one cares about. Call it $10 or maybe $50. If the site is causing damages to you, the $10 is worth it, but if they aren't causing the $10 of damage to make it worth filing then nobody else should have to deal with that paperwork either.

The DMCA does not give service providers leeway to check DMCA claims. Ignoring a correctly filed notification, even if it is garbage, prevents a service provider getting the relief from liability.

Source: 17 USC 512(c)

And "Big Copyright" prefers it that way. They are only interested in maximizing sensitivity (probability of taking down a true violation), they have no incentive to care about the trade-off with precision (probability that a taken-down violation is a true violation).

They process incorrectly filed notices too, for example notices that don't contain the information required. I had my content taken down once in response to a notice that just had a person's name and no other information (they were not the rightsholder)

The service provider is only liable if an infraction has actually taken place. Checking DMCA claims means determining if that is the case. Service providers do have that leeway.