August 08, 2012

Mars vs. Copyright vs. YouTube

As NASA's Curiosity probe made a spectacular and picture-perfect landing on the red planet a few days ago, much of the world was gathered around their televisions and computer screens to see the live show (luckily, since the interplanetary performance wasn't limited to NBC, we didn't have to wait for a prime-time delayed presentation).

In the wake of the landing, vast amounts of associated NASA-generated video -- all of it in the public domain -- has been incorporated into shows and other videos being uploaded by users of Google's YouTube.

But the incredible popularity of this new NASA footage has once again spotlighted an enduring issue with YouTube, which in general has been a matter of increasing concern and irritation for many YouTube users.

As we've discussed previously, the rights issues surrounding videos on YouTube can be very complex, as is the history of YouTube itself -- not to mention ongoing legal actions such as Viacom's massive (and many would argue, significantly hypocritical) associated lawsuit against YouTube/Google.

Google employs a number of mechanisms to try deal with these complicated intellectual property concerns, relying mainly on automated systems to process the vast amount of video uploaded to YouTube continuously -- almost 8 years worth of video every day. To instead pre-screen and pre-approve such a mountain of materials with human screeners would obviously not only be impractical, but raise the specter of outright censorship as well.

There are basically two enforcement systems on YouTube. One is "copyright strikes." These are the really serious ones. A strike (a specific claim by a video owner that material is directly infringing) results in the immediate removal of the video in question, and three strikes can result in the removal of all the user's videos and closure of their account. Strikes can take about six months of "good behavior" to expire.

Much more common, and by virtue of this fact probably far more irritating to far more users, are Content ID (CID) hits. These are generally fully automated events on YouTube (and on Google+ "Live On Air" Hangouts as well), triggered by pattern content matches against materials and resulting templates provided by participants in YouTube's CID program.

CID hits don't impact the overall status of user accounts. There is not (as far as I know) any limit to the number of CID hits an account may accumulate.

But CID hits do matter, and they can be maddening.

Because they are fully automated, "false" CID hits are extremely common. In same cases, these are the result of overzealous (or even fraudulent) CID users.

There can be an incentive for such behavior. The "owner" of a CID claim can effectively choose the penalty -- which can then be arbitrarily changed at any time. This can range from blocking of the video globally, in certain countries or regions, or -- more commonly -- the displaying of ads along with the video, with revenue going to the CID claimant, not the party who uploaded the video.

Sometimes this CID ad model works out well. It has permitted many YouTube users to use music in their videos without appropriate licenses, often subject only to a "Buy the music here" or other ad-related insertions.

But if a video uploader feels that they are using all public domain materials, or materials they created solely by themselves, even the presence of outside ads can be irritating -- and obviously it's much worse when video blocking due to CID claims is in play.

The CID "video reuse" problem is a classic example of how things can go wrong, and one that people ask me about in relation to Google nearly every day.

For example, if party A (not a CID participant) uploads a video that contains some public domain or original material, and party B (a CID participant) uploads a video that also uses some of the same public domain content (or even some of party A's original content), CID may in some cases flag party A's video as being in violation of party B.

I've had this happen myself. I've had both public domain video and public domain music excerpts falsely flagged by CID.

And this is also what's been happening with the new NASA Curiosity footage. Various ordinary YouTube users who displayed that public domain material in their uploaded videos are in some cases reportedly being swamped by Content ID hits from CID participants like news agencies and the like, who also used that same public domain footage. This has even happened to NASA itself on its own YouTube channels with its own videos.

When a Copyright Strike or Content ID hit occurs, YouTube sends the associated user an alert and pointer to forms for disputing the claim. These forms have been significantly improved recently, becoming easier to understand and covering more common use cases -- though there are still situations that can be difficult to accurately characterize via the forms.

Typically, when you file a dispute -- e.g. "The material in question is public domain." -- videos are essentially immediately restored and outside ads pulled down, until the claimant responds to the dispute.

Sometimes this process flows reasonably. Other times, claimants seem to perform what amount to pro forma rejections of disputes (returning the videos to their blocked/ad-insertion status), leaving the typical uploader with no practical further recourse.

Fundamental to many person's complaints about this situation is the implied "guilty until proven innocent" nature of the system. Even with the dispute mechanism in place, this strikes many observers as fundamentally unfair.

On the other hand, Content ID is fundamental to letting people upload video at all without pre-screening authorization. Trying to handle this primarily without the use of automated systems would be unthinkable.

Answers to these dilemmas don't come easily. You can't simply assume that the first party to upload material has the rights to it.

But it could be useful for parties who generate a great deal of public domain video (like NASA) to have an easy way to quickly indicate the status of their videos, so that included footage would not be used as triggers for CID hits.

A bit more "friction" in the system when dealing with materials that are not already known to be widely abused could also be useful.

In such cases, instead of blocking/ad-insertions immediately applying upon the triggering of CID, perhaps a warning could be sent out to both parties allowing them to discuss/negotiate the status of the footage in question (say up to 24 hours or some such) before primary CID-triggered actions (such as ads/blocking) would be deployed.

Some might argue that this would create a "window" for purposeful uploader abuse, but I believe this could still be limited algorithmically. Another counter-argument could perhaps be that busy media firms with large video holdings might not always be able to respond within a short time frame.

All that said, the underlying YouTube "penalty first, then try prove your innocence" status quo, even given the existing dispute mechanisms, seems increasingly untenable.

Legitimate rights holders most certainly should be able to protect their content.

But we should also be able to do better all around to create a more equitable framework for dealing with these situations, which would include the automated systems, routine dispute processes, and a means for escalated appeals.


Posted by Lauren at August 8, 2012 10:19 AM | Permalink
Twitter: @laurenweinstein
Google+: Lauren Weinstein