Monday, April 16, 2007

Automation in Takedowns

In my last post I linked to a summary from a Fordham conference where lawyers from Viacom and Google/YouTube presented their arguments at a panel. Those into these kinds of things know the arguments on both sides, but it's still nice to see these guys commenting on the thing in public. Notably, the summary says this about Alex MacGillivray (the GooTube rep) and his response to the accusation that YouTube's business model is based on infringing copyright:
Macgillivray then states that if YouTube were trying to make a business off of copyright infringement, it wouldn’t be doing all that’s doing now to curb infringement. First, he claims that they have the best DMCA notice/takedown procedure online — the turnaround is quick and automated.
Wait, what? An automated notice and takedown procedure?

I must admit, and am embarrassed to say so, that an automated notice/takedown provision is not something that had ever occurred to me. It strikes me as obvious now. It also raises some interesting questions about the DMCA, automating legal processes, and the Viacom/YouTube spat.

First, the provisions on the notice and takedown procedure in DMCA 512(c)(3) are ideal for being handled by an automated process. A copyright owner need provide an ISP with a description of the owner's work, an identification of the alleged infringing work (a URL normally suffices), as well as an address and statements saying that they own the copyright and that the use is infringing. As far as an ISP is concerned, the URL is all that matters in the notice as identifying the URL and disabling access to it can be easily automated and free up employees from handling these tedious and repetitious tasks.

Not only is there a standard form for DMCA notices, but the incentive structure of the DMCA further encourages automation. DMCA 512(c) says that an ISP will not be held liable for monetary, injunctive, or other equitable relief for copyright infringement so long as the ISP "responds expeditiously to remove, or disable access to, the material." Further, under 512(g), the ISP is also immune for disabling access to material pursuant to an illegitimate takedown notice. Thus, an ISP is encouraged to takedown videos at a copyright owner's say-so and are immune from suit if it turns out the copyright owner is abusing copyright.

It is obviously more efficient if YouTube and the like can automate this notice and takedown process, rather than pay people to read over these letters and manually remove the videos. Based on the structure of the DMCA, there's no doubt that a human would decide, or be ordered to by a boss, to behave in the same way and to disable access to any material that is identified in a takedown notice. If the takedown notice is flawed in some way, either because the notice wasn't sent by the copyright owner or because the material is outside of copyright (eg. fair use), that's not the ISP's problem. The user can request to have the material restored or can bring the issue up directly with the alleged copyright owner. There's nothing for the ISP to do but take the material down.

This automatic removal of material seems troublesome in this context because of the interest in free speech. People are allowed to make a fair use of another's copyrighted content and depriving people of that ability raises a concern about people being able to freely express themselves. I refer you to the EFF's suit against Viacom for removing a YouTube video that made use of some Colbert Report images to poke fun at the icon and If a person with rudimentary knowledge about copyright law saw that takedown notice and then watched the video, it is highly possible the video would have stayed and the lawsuit could have been avoided. Instead, this funny video that satirizes current political and entertainment figures was essentially censored.

And it's not just the takedown notices used to push the bounds of what is covered by copyright that cause problems. Just this weekend it was reported that a 15 year-old sent takedown notices to YouTube to remove Australian Broadcast Company videos. The kid claimed in his notice that he represented ABC and, of course, the authorized videos were disabled. Might a human agent be able to make sure takedown notices are actually coming from the entities they claim? Is the cost of doing so worth it for an ISP like YouTube given they have no incentive to do so and face no liability for immediately removing the video?

Granted we are talking about video clips, a few minutes long, that may be of low social or personal significance. But this kind of automated legal process, where legal rights are determined by algorithm and no human intervention, can potentially reshape copyright law and its volumes of intricate analysis meant to draw lines between acceptable and unacceptable behavior. It's the privatization of law, taking it away from neutral magistrates entrusted with applying the law. The automation of YouTube's notice and takedown process may just be an isolated case of using technology to create efficiency with a minimal loss of rights, but a slippery slope argument could be made that lots of entities would like to see these same kinds of efficiencies (notably roving internet wiretaps and data-mining that can be used to trigger criminal investigations).

Whether the sky is falling is not for this post, but the issues raised by the automation of the notice and takedown process does have implications for the Viacom/YouTube lawsuit.

YouTube's hands off approach to these notices highlights the fact that the dispute over unauthorized videos really lies between copyright owners and the videos themselves. There's a reason why the DMCA was structured as it is to remove from ISPs the burden of making legal judgments about whether a takedown notice is legit or not. It's a tough call (could be fair use) and inefficient to thoroughly research (could be a 15 year-old kid or my dog sending the notice, how much money should an ISP spend to find out). If removing this burden was important to the drafters of the DMCA, then why should YouTube be burdened with monitoring for copyrighted videos (essentially the same as making legal judgments on takedowns) when the DMCA pretty much says that the notice and takedown procedure is a copyright owner's only avenue for redress?

Because the DMCA strives to turn an ISP like YouTube into a mere repository of user submitted videos, it's pretty clear that lawsuits should be going from copyright owners directly to the users uploading the videos. The structure of the DMCA encourages this result: the copyright owners sends a takedown notice to YouTube, who removes the video; if the user objects, a counternotice is sent to YouTube; if the copyright owner doesn't object to the counternotice the video goes back up, if it does object, the parties end up in Federal Court. In the end, the DMCA gets the copyright owner and the user into court and leaves YouTube alone to continue offering its service.

In short, that YouTube has an automated notice and takedown process shows that YouTube is avoiding making legal determinations about the videos on its site. This process could only be automated if the DMCA was so structured as to make it a reasonable thing to do. That the DMCA, by its words, aims to remove the burden of making these legal determinations from YouTube and works to bring the copyright owner and users into court would suggest that YouTube is following the DMCA in respect to Viacom's videos. If Viacom has issues with YouTube, those issues could be better characterized as being with the DMCA itself. Viacom's own interests in protecting its business model and keeping its big brand advertisers happy (by preserving their business model of controlling the context of their ads) almost require that it do something to change the video landscape. Viacom would be better off lobbying for new legislation than suing YouTube, which appears to be following a law that basically orders it to behave this way or get sued to death. Oops.
View blog reactions

AddThis Social Bookmark Button


Anonymous Digital 3d Art said...

I have a case of Googles automated process I am sure of being used on bloggger. I feel that they have abused safe harbor in that they were removing posts that included digital artwork I did. Then the algorythm changed and now the system is not removing my work and telling me to take it up and with and anonymous offender. This is and offender that is hiding behind the way the blogger platform works.

12:41 PM  

Post a Comment

Links to this post:

Create a Link

<< Home