March 21, 2016

Google Questions & Unofficial Answers: "Why does Google's YouTube seem so biased against ordinary users who upload videos?"

(This is a new entry from my recently formed Google+ Community "Google Questions & Unofficial Answers" -- located at: https://plus.google.com/communities/111481685962796128662. )

Why does Google's YouTube seem so biased against ordinary users who upload videos? I've unfairly had my videos blocked, received copyright strikes for my own materials, and even had my account suspended -- and it's impossible to reach anyone at YouTube to complain!

No, YouTube isn't biased against you -- not voluntarily, anyway. But it could definitely be argued that the copyright legal landscape -- particularly in the mainstream entertainment industry -- is indeed biased against the "little guys," and Google's YouTube must obey the laws as written. What's more, YouTube exists at the "bleeding edge" of the intersection of technology and law, where there's oh so much that goes bump in the night.

Let's begin with a fact. The amount of video being uploaded by users around the globe into YouTube at any given moment is staggering. As of July 2015 last year, something like 400 hours of video were being uploaded every minute (!), up from 300 the previous November. You can only imagine how much is pouring in today. That's one hell of a lot of video.

When we talk about uploaded videos, it's not just Internet bandwidth and disk space, it's also processing such as transcoding, sorting, analysis, and much more -- a whole array of activities triggered by every single "simple" YouTube upload.

At these kinds of data volume levels, pretty much everything has to be entirely automated for the overwhelmingly vast majority of videos. Manual processing, or manual responses to every or even most user queries or complaints, would be utterly impractical.

Obviously, money is an important aspect of YouTube. Content owners can earn revenue from user views of their content via ads, and Google generates income in the process. Since there are crooks around attempting to game that system (e.g., through false clicks and fake views), significant resources must be devoted to detecting and eliminating their impact as well. And YouTube operations don't come cheap. Outside of the uploading numbers above, think about all the people using YouTube-related resources to view videos at any given moment around the world. YouTube has over a billion users. Hundreds of millions of video hours are viewed via billions of YouTube clicks every day! And yes, Google wants to quite appropriately make a profit with YouTube as well.

This brings us to the real heart of the matter, where brilliant YouTube engineering meets The Twilight Zone -- in other words (drum roll, please): the legal system.

Here is a truism that may give you a headache to even think about: Many of the key aspects of YouTube that ordinary video uploaders consider to be the most bizarre and unfair are fundamental requirements to helping make YouTube possible at all!

Without YouTube's Content ID system that permits content owners to detect and monetize material they own that YouTube users have uploaded without permission or rights (e.g. popular music clips, to name but one of many examples), the likely outcome in the vast majority of cases would be complete takedowns under the DMCA (Digital Millennium Copyright Act) and other laws -- again, all of which Google/YouTube must comply.

A big plus for all of us from Content ID -- and key to keeping so many great videos available on YouTube -- is that it provides content owners with alternative options to total takedowns -- such as blocking only in certain geographic areas, monetization by the content owner rather than the unauthorized uploader, and so on. Similarly, the YouTube "three strikes" copyright violations policy, and other related Terms of Use policies, are themselves alternatives to the otherwise "most likely under the law" outcomes of immediate account terminations and even legal actions being taken against unauthorized uploaders by content owners.

None of this is to suggest that everything is butterflies and rainbows with Content ID. Like any system -- especially one that rides the thin line between technology and the volatile world of courts and lawyers -- it is not a perfect mechanism.

Crooks are continually trying to circumvent Content ID, to monetize videos over which they have no rights at all. I've made a sort of a hobby (yeah, I have some eclectic hobbies) of watching for these and reporting them to YouTube, but they're fairly easy to find. Just do a search on YouTube for pretty much any well known movie you've ever heard of, or most popular television shows even many decades old. Odds are you'll get lots of results, many of them seeming incredibly recent (like uploaded only a day or even just a few hours ago). Their large quantity suggests automated systems doing the uploading, usually to short-term "throw-away" accounts. If you actually try to view these videos, you'll typically find they're either nothing but raw spam -- displaying a link urging you to go to pirate sites to see the actual videos (where you're likely to be met with dubious credit card requests, malware, or worse), or monetized versions of the films or shows that have been altered in ways to try evade Content ID for as long as possible (the methods employed range from comparatively subtle, to horrific and bizarre visual distortions).

I hate these kinds of outright cheaters. They're trying to manipulate users into viewing spam and/or substandard perversions of the original films or programs, to try make money from content over which they have zero rights. YouTube is constantly working to fight them, but as a fan of classic movies and TV -- and of YouTube -- I personally feel that this category of copyright violators deserve no leniency.

The flip side is that there are situations where innocent users can become inappropriately targeted by Content ID or YouTube's copyright strikes reporting systems, via false positive Content ID hits, inappropriate copyright claims, and associated video demonetizations, takedowns, and account suspension/termination actions.

False claims against YouTube videos by content owners (or purported content owners), either purposefully and accidentally, both by design and sloppiness, occur every day. At YouTube scale, significant numbers of users are affected.

Such situations can get pretty "meta" too. There are all sorts of complexities surrounding figuring out what is actually "public domain" video, and how to deal with it. For example, think about the case of a content owner who uses public domain material in their own production, who then inappropriately claims rights against a third-party production that happened to use the same public domain clip as that claiming party. There are also cases of content claimants claiming the rights to materials completely produced by someone else, when that original material was partially or wholly incorporated into a larger production by the claimant. Your head spinning yet?

The concept of "fair use" -- tough even for the courts to deal with over the years -- is currently very difficult to incorporate in a useful form into scanning algorithms. Classical music has been a traditional problem as well. I personally know one YouTube user who performs long classical pieces on the piano, who has repeatedly had his YouTube videos demonetized because Content ID was trying to incorrectly claim his performances for other parties (this is a tough kind of case, because high quality performances of the same classical piano composition performed by two different excellent players can sound very similar). The poor guy actually was resorting to purposely incorporating errors into his piano recordings to try differentiate them when uploaded. Fortunately, YouTube has been making considerable technical strides in minimizing the problems that have affected him and other users related to these kinds of analysis.

Yes, when false claims or other similar problems hit an ordinary uploader's YouTube video, it could indeed seem like a Kafkaesque, automated forms ordeal to try resolve them. This situation is improving -- YouTube has actually been making dramatic improvements in their claim/counterclaim resolution flow -- although some problems in these respects still definitely persist.

Keep in mind -- as was noted earlier -- that at these video upload volume levels, most or all stages of the process must by definition involve automated rather than human-based analysis, but also crucially, the DMCA and other related laws impose an extremely limited range of options with which YouTube can deal with these situations and stay within those laws as YouTube must -- even in some cases when faced with abusers of the DMCA who make repeated false content claims.

Google knows there's a lot more work to do in this context. YouTube last month publicly announced ( https://productforums.google.com/forum/#!topic/youtube/x3aGmn_MsqI ) the creation of a new team specifically to improve transparency, communications, and associated processes across a range of these issues. What we also need is reform of the entire copyright ecosystem to more fairly treat ordinary users instead of the "guilty until proven innocent" skew that current content ownership/copyright laws tend to require -- though given our current toxic political environment I wouldn't bet the farm on the likelihood of positive legal changes in this regard anytime soon.

The various Google/YouTube teams who breathe this stuff 24/7/365 try very hard to get it all right. But when it comes to video and the Internet, especially when one considers the multitude of complicated, multidisciplinary aspects, nothing is trivial nor comes easily in the associated technical, policy, or legal realms.

Be seeing you.

--Lauren--
I have consulted to Google, but I am not currently doing so -- my opinions expressed here are mine alone.

Posted by Lauren at March 21, 2016 09:03 AM | Permalink
Twitter: @laurenweinstein
Google+: Lauren Weinstein