Archive

Posts Tagged ‘Copyright Directive’

The Ennui of Learned Helplessness: Article 13 and the Five Lies in YouTube’s Content ID

March 29, 2019 Comments off

youtube-logo-parody-1

According to Wired (“Don’t believe the hype: Article 13 is great news for YouTube“), YouTube is positioned to be a big winner due to the Article 13 requirement for “upload filters”.  If you’re keeping your brackets for “Most Googlely Journalist” in the post-Article 13 March Madness spin, Wired gets the three point play on this post–there are no upload filters in Article 13, so not quite sure what Wired is getting at here.  But I digress.

Wired tells us:

[Article 13’s upload filters are] likely to be disruptive, for YouTube as for everyone else. There will be mistakes, disgruntled creators and meme-posters, protests against this or that algorithmic decision. But if anyone is going to eventually benefit from this, in the long run, that is YouTube.

Leaving aside the fact that Alphabet, YouTube’s owner, would probably have enough money to invest in licensing agreements with media companies and music labels – YouTube is uniquely positioned to capitalise on the internet’s sudden need for copyright filters. The platform itself has been using a type of copyright filter, its home-brewed Content ID, for over a decade. The algorithm compares newly uploaded footage against a database of registered videos, and demonetises (or takes down) any post containing matching content [which is also not quite right–it allows the user of Content ID to elect to block or monetize, with heavy pressure to monetize and not block].

While there have been instances of manipulation and egregious mistakes, Alphabet has invested more than $100 million in Content ID’s development, and the technology is already used by more than 9,000 broadcasters, movie studios, and music producers globally. [And there it is–we’ll come back to that] A vast number of other companies – virtually, every single platform and website that does not want to fall foul of Article 13 – could soon swell the ranks of Content ID’s users. Rather than an existential threat, Article 13 could wind up being a fillip to YouTube’s finances.

First, let’s dispense with an implication of that last assertion–that YouTube would profit from ContentID by licensing Content ID to third parties.  There is absolutely no evidence that YouTube would do anything remotely like that, and even if they did, Google would likely still control who gets to put their works into ContentID in the first place.  Wired doesn’t actually say this, but a reader might get that implication.

What is more clear is that Wired asserts YouTube would have a competitive advantage over other, smaller perhaps, platforms–not like they already do as the result of Google’s illegal favoring of its own products for which it is being fined billions in Europe.  This due to the nonexistent “upload filters” that are not in the Copyright Directive (aka Article 13) but that the tech press keeps asserting are really there in one of the great gaslighting exercises of all time.

Remember–ContentID has a lot more to do with preserving YouTube’s US-based DMCA safe harbor and getting licenses for premium content (with higher advertising revenue, i.e., CPMs) than it does with some desire to do the right thing.  That sentiment arguably does not exist at Google, YouTube, Facebook or any other Silicon Valley company with the notable exception of Apple.

OLYMPUS DIGITAL CAMERA

Wired blows past the two key facts it its own story.  First, the number of participants in ContentID:  “the technology is already used by more than 9,000 broadcasters, movie studios, and music producers globally”.

This statistic comes from YouTube itself:

Number of Content ID Users

But strangely according to a 2016 story in the New York Times:

YouTube says that about 8,000 companies and organizations have access to Content ID and that independents may get access through affiliated companies and industry groups.

That’s right–the number of users of ContentID has increased–worldwide–by 1,000 in four years.  Any idea how many new videos were uploaded to YouTube in that time?  We may not have that exact number, but we do know this again according to YouTube’s own statistics:

youtube number of views

That’s billion with a B.  So just rough justice, don’t you think that if there are that many videos being viewed on YouTube there would be more than 9,000 worldwide users of ContentID?

The other relevant fact is that Wired breathlessly repeats that YouTube spent $100,000,000 on developing Content ID.  According to Wikipedia (which I tend to believe in this case because it’s their benefactor Google), ContentID cost $60 million to develop by 2016 and as of 2018 Google had spent $100 million on the system.

Allow me to posit that $100 million for a system that can handle the volume on YouTube is chump change.  One reason that it cost so little is that it is working for purpose–it is not intended to catch everything, it is only intended to catch works by the people who sign YouTube’s chump deal or people who are “important” (in the best traditions of YouTube’s founders).

Remember this line from the 2016 NYT story?  “[I]ndependents may get access [to ContentID] through affiliated companies and industry groups.”

So you mean that some artists are more equal than others?

Exactly.

YouTube’s theory according to the NYT is that independent artists (such as five time Grammy-winner Maria Schneider who graced our pages with her groundbreaking essay on YouTube’s sleaze) are not harmed by YouTube’s “catch me if you can” DMCA shakedown because Content ID is widely available.  The implication being if those pesky artists would just use the tools YouTube provides, there would be peace in the valley with sunshine, gum drops and puppy dog tails for everyone with happiness among the subjects of the Unicorn Kings.

The clear implication is that “independents” have nothing to complain about because they can get “access” to ContentID through “affiliated companies and industry groups”.  “Affiliated” in this case means affiliated with YouTube (laughably called “partners”), and that means that the “companies and industry groups” have signed a ContentID license agreement which is essentially a nonnegotiable form contract imposed by Google.

Because Google wants to have the rights to use their IP all tied down.

Ahem.

This is another reason why Wired should not write to the Google press release.

So let’s start with what YouTube actually says about who gets ContentID:

Content ID UseAnd what are the “specific criteria” that copyright owners have to meet for their “substantial body of original material”?

qualifications content id

So that quote from YouTube’s website arguably explains why there’s only 9,000 entities that have access to Content ID on a worldwide basis across all copyright categories (assuming that’s even true).

There’s at least five lies underlying Content ID, all of which you’d miss if you didn’t have the inside baseball insight into the unnecessarily complex Content ID system–and as we know, complexity almost always hides fraud.

Lie #1: Show Me Where I Signed Your Social Contract

The first point is why should artists be required to even deal with Content ID or YouTube at all?  If an artist never consented to being on the site in the first place, why should Google be able to just exploit their work without consent?  Why shouldn’t Google have to have a contract with the artists to exploit their IP?  You know, the way you have to be approved and have a license to use Content ID.  This is a core property rights concept underlying the new Copyright Directive in Europe–we hope that the US follows suit with DMCA reform.

There is tremendous cost associated with engaging with YouTube at all whether you qualify for Content ID or  you don’t.  In fact, as the Trichordist’s Streaming Price Bible demonstrates, YouTube’s royalties are so crappy that it’s entirely possible that the total cost of doing business with YouTube exceeds any royalties you could make–because the cost of dealing with YouTube varies directly with the size of your catalog.

So why shouldn’t artists be able to just say no and keep all of their music (or other work product) off of YouTube?  Every penny spent trying to block unauthorized videos is a penny spent for YouTube’s benefit.  And why is it we have to pay for this?

The truth is that it is not at all apparent that declining the opportunity to license YouTube wouldn’t actually be more profitable than dealing with the incredibly screwed up Content ID and CMS system.

So let’s not assume that Content ID and notice and shakedown are the only possible outcomes here.

Lie #2: Using Content ID Is Not Free

Even though Google doesn’t charge for Content ID, using the system is hardly free, especially for “independents”.  In order to get “access” to Content ID, an independent artist needs to contract with a claiming company–and pay that company anywhere from 20% to 50% of their YouTube revenue.

And let’s be clear–claiming companies exist to fix YouTube’s mistakes imposed on the world due to YouTube’s legacy and highly inefficient DMCA notice and shakedown business.  Every penny spent by an artist through giving a claiming company a revenue share is a penny spent for YouTube’s benefit by an artist capitulating to the notice and shakedown onslaught.

So saying that “independents” have “access” to Content ID through a claiming company “affiliated” with YouTube is a grotesque oversimplification.  There are claiming companies that operate at the more lucrative end of the YouTube doing channel management and MCN or near-MCN business for which they may operate their own in-house advertising sales staff.

The claiming companies in reach of “independents” necessarily have to take a larger share of a smaller revenue stream in order to operate.  And here’s what they don’t do:

Block.

Why do they only monetize?  Because that’s what a revenue share means–revenue.  Using Content ID to just block videos (especially UGC) would only be available for a fee (since there’s no revenue if you block everything).  Independent artists can’t afford to pay a fee to block on YouTube so they typically will capitulate and monetize.

And who benefits from that?  YouTube.

Why would blocking require a fee for service?  If an artist just wants to bail out altogether, then that artist would set the automated controls of CMS to block worldwide.  In order to make that blocking meaningful, there would need to be a lot of manual care and feeding to account for UGC leakage through the very porous Content ID.

That would include techniques like pitch bending to use the curious speed controls on the YouTube player which seem to have one purpose–defeating Content ID.

This is what’s called a royal pain in the trade, so anyone doing that work would have to be paid for the hours and hours and hours it would take to accomplish it.  Since the artist can’t afford to pay someone else to do that work, the artist would need to do it in all their spare time.  Which of course will not be very effective or may not happen at all.

I call this the ennui of learned helplessness.

Lie #3: Artists Cannot Access Content ID

By using the word “access” when it comes to Content ID, YouTube is equivocating yet again.   If you are an independent artist and your distributor has a CMS account (and that’s a small group), do you have access to Content ID?

No.  At best, you can tell your distributor what you do and do not want monetized.  They will only devote so much time to you, however, and they won’t do the manual claiming on UGC, etc., at least not until you get some pretty significant traction on YouTube (meaning over 5,000 views or so on a particular video).

Your distributor will not allow you to get your hands on their CMS or Content ID dashboards.  There’s a good reason for this, which is that the way Google licenses Content ID there’s a good chance that the distributor (such as Tunecore or CD Baby) could never get enough seats for its particular CMS license to allow all the distributed artists to have individual access, and there’s no view in Content ID that would show one artist’s tracks without showing that user all the other artist’s tracks handled by that distributor.

Why?  Because YouTube doesn’t design the system for “independents”.

Lie #4:  Independent Songwriters are SOL

Notice YouTube never talks about independent songwriters having “access” to Content ID.    The closest that an independent songwriter comes to getting access to Content ID is if they opted into the HFA YouTube license connected to the out of court settlement of the class action against YouTube that was a companion case to Viacom v. YouTube (and which wasn’t certified as a class, but is often referred to as a class action by people wishing to avoid using the legal term “putative”).

So ask independent songwriters who opted in to the HFA license how that “access” is working out for them.

Lie #5:  Content ID Is Another Nondisplay Use of Other People’s Stuff

Google has made a subspecialty of acquiring data for one use and actually using it for other purposes–undisclosed purposes.

Remember “GOOG-411”?  This Google product was the “free” Google directory assistance (very similar to Google Voice). Former Googler (and former Yahoo!er) Marissa Meyer told  Info World years ago that GOOG-411 was not intended to be what it appeared to be:

You may have heard about our [directory assistance] 1-800-GOOG-411 service. Whether or not free 411 is a profitable business unto itself is yet to be seen. I myself am somewhat skeptical. The reason we really did it is because we need to build a great speech-to-text model … that we can use for all kinds of different things, including video search.

The speech recognition experts that we have say: If you want us to build a really robust speech model, we need a lot of phonemes, which is a syllable as spoken by a particular voice with a particular intonation. So we need a lot of people talking, saying things so that we can ultimately train off of that. … So 1-800-GOOG-411 is about that: Getting a bunch of different speech samples so that when you call up or we’re trying to get the voice out of video [such as from YouTube], we can do it with high accuracy.

That’s right–Google told you the product was doing one thing, but in actual fact it was always intended to be something entirely different.  The real action was in the background where users couldn’t see it.  If Marissa Meyer hadn’t let it slip in an interview, you might never have known.

If you have a Content ID contract, check out this language in paragraph 2:

By providing Reference Files, you grant Google a non-exclusive, royalty-free, limited license to (a) store, copy (including the right to make temporary cache and storage copies), modify or reformat, excerpt, analyze, use to create algorithms and binary representations, create ID Files and otherwise use those Reference Files, the ID Files and the associated metadata in connection with the System

And there it is:  “otherwise use”.  Pretty broad grant of rights, eh?  You could say that “in connection with the System” is limiting, but how would you ever know what “in connection with” means?

Remember Google Books?  Ever heard of “corpus machine translation“?  Google uses the scans of the tens of millions of books it stole from authors in the background to improve its translation algorithms.  If the authors had brought their case about that, do you think the court would have been so quick to find this obviously massively commercial application a fair use?

Bevo ≠ Unicorn King

Once again, YouTube has scammed their way past artist objections such as those in Maria Schneider’s post and Irving Azoff’s open letter.  I think this is partly because the whole Content ID system is such inside baseball–once you accept the idea that requiring artists to use these legacy DMCA tools is even acceptable, which I don’t.  Reporters just don’t know what questions to ask.

Now they do.

(Some of this post previously appeared on MTP in other posts.)

Goliath Never Learns: “Article 13” Passes European Parliament With Substantial Majority–now what?

March 26, 2019 Comments off

The European Copyright Directive (aka “Article 13”) has passed the European Parliament by a vote of 348 to 274, a substantial majority in favor of artist rights.

Final Vote Article 13

Silicon Valley interests embarrassed themselves as expected.  Throughout the legislative process, Google in particular demonstrated a yearning for mob rule and fielded well-financed fakery both online and offline as confirmed by The Times of London (“Google funds website that spams for its causes”),  several articles in Frankfurter Allgemeine Zeitung, research by David Lowery and Volker Rieck, and on MTP.

Dj_qcYOW4AAfyRy.jpg-large

YouTube was caught red-handed by FAZ funding YouTubers who created messaging adhering to the party line against Article 13 and actually managed to create a backlash by YouTubers who thought the Google tactics reeked, such as the French group Tatou:

And it must be said that in the middle of all of this mess over the Copyright Directive, Google was fined not once but twice by the European Commission for a total of over $2 billion for blatant competition law violations.  So not only was Google improperly exploiting its monopoly position commercially for which it was fined heavily, it was simultaneously exploiting its monopoly position to manipulate users for which it has yet to be punished.

Punishing Google for message manipulation is not an idle thought–Google was intentionally directing its lobbying message at children–we know this because the children of MEPs and artists told their parents that copyright was evil courtesy of YouTube.

32ed61f04626eaad138c7a4e01f713ce

Another part of the disinformation campaign was Google’s flood the zone strategy using bots on social media–which is not yet proven conclusively, but is the kind of thing that people with badges are good at tracking down.  Here’s a little taste of what this bot campaign looked like.

As one journalist noted, “I’ve concluded that there are only three people in Europe objecting to Article 13, but one of them draws the avatar for 500 trillion bots.”

The typical criteria for discovering bots are attributes like follower/following totals out of sync with number of likes, such as this:

Bot 8 2-12-19 Edited

Or this:

Bot 3 1-5-19

In the words of that great philosopher Chico Marx, ya gonna believe me or your lying eyes?  If you really believe that this bot activity sprang up all by itself and had nothing to do with Google even though the bots parroted the Google line, I’m sorry–I just can’t help you.  Try searching Twitter for #article13 and see what comes back.

And then there were the coordinated attacks on MEPs through spam and telephone.  Millions of emails over the last year, the overwhelming majority of which probably did not come from Europe.   Not just mere fakery, but fakery created by a multinational corporation to further its commercial interests through a massive disinformation campaign.

And of course the main reason that the disinformation campaign was so massive is because of Google’s monopoly position in Europe–the very thing that the European Commission is fining them billions for abusing.

This is not mere speculation–Google and Facebook told MEP Helga Truepel to her face that they intended to interfere in the European legislative process, and faster than you can say Internet Research Agency, they did.

helga google interfere

It also must be said that the election campaigning by Google against the MEPs who dared to vote against Silicon Valley suggests that the EU election transparency laws may well have been violated perhaps by sitting MEPs.  These were not spontaneous actions, they were part of a calculated campaign.

Bots 14-2-19 Edited

The question now remains–what is the European Parliament prepared to do about it?  Someone needs to do something because one thing is clear, they will be back.  The illegal tactics will be even worse next time because Goliath never learns.

 

 

A Cautionary Tale for Congress: Big Tech Uses Fake Grassroots Interference Lobbying In Europe

March 11, 2019 Comments off

There’s a sound policymaking reason why the European Parliament should ignore the bombardment of email and social media messaging it has recently endured for copyright reform.  That reason is well articulated in a 2010 memo by Professor Cass Sunstein (then Administrator of the White House Office of Information and Regulatory Affairs) when he cautioned the Obama Administration against relying on social media for making policy:

“Because, in general, the results of online rankings, ratings, and tagging (e.g., number of votes or top rank) are not statistically generalizable, they should not be used as the basis for policy or planning.”

In a post-Cambridge Analytica world where social media platforms not only are ubiquitous but also have earned the distrust of policy-makers and voters as well as parents, Sunstein’s nine-year old admonition is especially prescient-but he could have added “easy to fake.” This would apply to the new boiler rooms of fake Twitter accounts with distorted ratios of tweets to followers, or followed accounts to likes–such as the infamous Internet Research Agency.

Sunstein’s memo is both ominous and also relevant given developments at the European Parliament over the last few months.  Big Tech’s interference lobbyists are fighting the last war against safe harbor reform using “fake grassroots” online campaigns against Members instead of legitimate advocacy.

Specifically, the reaction of multinational Big Tech legacy players to the European Parliament’s safe harbor reform effort demonstrates a toxic brew of corporations attacking Members.   That’s a dark turn to Sunstein’s analysis.  Crucially, it appears that bot farming techniques and mass email attacks verging on denial of service are weapons in Big Tech’s interference lobbying arsenal despite their hollow assurances to multiple governments about policing interference by state actors in national elections.

Think it can’t happen here?  Think again.

Europe’s Safe Harbor Reform

Known formally as the “European Directive on Copyright in the Digital Single Market” or informally as “Article 13”, the EU legislation is intended to update the law and, among other things, substantially cuts back the legacy safe harbor privilege that is so valuable to Big Tech. (See Emmanuel Legrand’s excellent summary.)  It’s understandable that they would–Facebook, Google, Twitch and others have based their trillion-dollar market caps on what otherwise would be called piracy–also known as the “value gap”.

Congress has yet to grapple with the US version of the DMCA safe harbor (referring generally to Section 512 of the Copyright Act), although the Copyright Office is beginning field hearings on the much needed overhaul.  Originally conceived in 1998 as a way to afford a little latitude to reasonable people acting reasonably, Big Tech and its acolytes have distorted the DMCA “safe harbor” beyond recognition in the last twenty years and converted into an alibi.

In the last session, Congress took an important step by cutting back a different safe harbor in Section 230 of the Communications Decency Act with the SESTA-FOSTA package.  SESTA-FOSTA eliminated a safe harbor for sex trafficking which seems uncontroversial-but caused a huge lobbying counterattack by Big Tech.  If you caught any of the manipulated onslaught against SESTA-FOSTA, you have a taste of what the European Parliament is experiencing with their Copyright Directive.

There’s no question that Google and Facebook lobbyists interfered with the EU’s legislation–the companies threatened an MEP that they would interfere in the EU elections if they were challenged on Article 13:

German MEP Truepel gives a first hand confirmation of what happened.  Further, in an official blog post that has become controversial because it struck a little close to home, the European Commission commented conclusively on the MEPs’ experience with Big Tech’s “fake grassroots” campaign and interference lobbying tactics:

[T]here is ample evidence from respected sources, here and here and perhaps here or here or indeed here that ‘Big Technology’ has even ‘created’ grassroots campaigns against the Copyright Directive in order to make it look and sound as if the EU is acting against the ‘will of the people’….

Do Google, Facebook or others really need to pay to persuade?

Are we in a world where ordinary people side with the fire breathing dragon against the knight with a blue and yellow shield?

The Fake Interference Campaign

Yes, Google and Facebook attempted to pass off a blistering email spam campaign targeting Members as a spontaneous grassroots uprising.  Not just once, but three times on three separate votes in Parliament on the Copyright Directive-even after they had been caught faking and interfering by major news outlets.

The first “campaign” was around a committee vote that resulted in a win for Big Tech. Thanks to David Lowery and Volker Reick, the press called out the fakery in exposes by the Times of London and Frankfurter Allgemeine Zeitung among others.  Big Tech attempted to counter that expose by organizing in-person protests after making online threats of the “we’ll show you we’re real” variety across Europe.

Far from setting Europe ablaze, some estimate that these protests drew fewer than a total of 500 across Europe in multiple locations.  Pictures taken at these “protests” suggest that they were primarily organized by the Pirate Party, a fringe political party with one MEP that supports piracy and safe harbors as its name suggests.

Expose or no expose, the “fake grassroots” and spamming were repeated as the Copyright Directive went through successive procedural votes-which Big Tech lost miserably.  Yet the multinationals keep running the same interference play and kept getting sacked.

Members and the European Commission Speak Out Against Interference

Members were not fooled.  One example is an odd Tweetstorm against reform by suspicious accounts.  Members were subjected to attempts to stimulate the documented “emotional contagion” that drives social media.  A typical bot-like Twitter account against the Copyright Directive would have followers in low double digits or fewer but tens of thousands of tweets and a disproportionately high number of “likes” (in some cases over 100,000).  That ratio suggests that the account was used to capitalize on the Twitter “ratioing” algorithm to drive the corporate message into user timelines.  One wonders what did Twitter know and when did they know it?

In addition to the Tweetstorm, interference lobbyists ran a near-DDOS level email campaign that was actually counter-productive.  When asked at a press conference why the vote tally switched from opposing the Copyright Directive in the first committee vote to overwhelmingly supporting it in the next plenary vote, German MEP Helga Truepel pulled no punches:

“I think it’s due to this message spamming campaign. I talked to some of my colleagues here [and they] are totally [angry]…”

The European Commission’s official blog echoes Sunstein:

So next time, when you get a sponsored message on your timeline, which says something like ‘the EU will kill the world wide web as we know it’, stop, pause and consider for a moment. Ask yourself: Cui Bono? Who really benefits from this message or this wider negative campaign?

We know the answer to that question.  Google and Facebook became the biggest corporations in commercial history by manipulating legacy safe harbors in ways that neither Congress nor the EU intended.  Despite assurances about policing interference by bad actors on their networks, the European experience suggests they are actively using similar interference techniques to protect their privilege.

Europe has shown the world that legacy Big Tech business models can be brought into the 21st Century through well thought-out legislation.  But Congress would do well to arm itself against comparable “fake grassroots” interference campaigns in making policy for vital DMCA safe harbor reform to close the value gap.

[This post first appeared in the MusicTechPolicy newsletter, sign up for free version here.]

%d bloggers like this: