« Human "dignity" and relationships | Main | Why course packets ought to be fair use »

Thursday, August 28, 2008

Io v. Veoh: Harmful to YouTube?

Yesterday's decision in Io Group, Inc. v. Veoh Networks, Inc. is generally being hailed around the blogosphere as a win, or at least a bit of good news, for YouTube in its ongoing infringement fight against Viacom. See the L.A. Times, CNet, Wired, Silicon Alley Insider, Techdirt, TechCrunch. Certainly Google is trumpeting the case as a win.

The piece of the decision most commenters are focusing on is the holding that Veoh's automated reformatting of uploaded materials does not create copies for which Veoh is responsible. That's clearly a good precedent for ISPs that process uploaded content somehow (and follows on the heels of the Second Circuit's similar decision in the Cablevision case).

But there's a key component of the decision that is not at all helpful to YouTube.

The Io court specifically noted that Veoh was able to invoke the DMCA safe harbor, shielding it from liability, because "there is no indication that Veoh has failed to police its system to the fullest extent permitted by its architecture." (Slip op. at 29.) Among those policing efforts was Veoh's practice of using "digital fingerprint technology" to "prevent[ ] the some infringing content from ever being uploaded again." But as I've said before, I believe the real driver behind Viacom's suit to be its claim that YouTube has filtering technology that it refuses to use to police Viacom's content. From Paragraph 7 of the Viacom First Amended Complaint:

Moreover, YouTube has deliberately withheld the application of available copyright protection measures in order to coerce rights holders to grant it licenses on favorable terms. YouTube’s chief executive and cofounder Chad Hurley was quoted in the New York Times on February 3, 2007, as saying that YouTube has agreed to use filtering technology “to identify and possibly remove copyrighted material,” but only after YouTube obtains a license from the copyright owner.... Those who refuse to be coerced are subjected to continuing infringement.

So Viacom's claim is that YouTube is not policing its system to the fullest extent permitted by its architecture, and in fact is holding some tools back as a bargaining chip. To the extent Io and other decisions are holding that, to preserve its DMCA immunity, an ISP has to police its system to the best of its abilities, YouTube may be in trouble (assuming Viacom's facts are accurate).

It's important to note that in the Io decision, the implicit requirement to police the system comes up in the context of discussing vicarious liability, which is not immunized under the DMCA safe harbor. And vicarious liability has two prongs to it, (1) the "right and ability to supervise the infringing activity" and (2) a "direct financial interest" in those activities. The Io court's discussion of policing occurred in the context of addressing the first prong. The second prong would still need to be demonstrated, and that would likely require a showing that, e.g., YouTube's ability to sell ads was somehow enhanced by the infringing content. Maybe that would succeed, maybe it wouldn't.

But I think the same "duty to police" theory may apply under the core of the safe harbor itself, Section 512(c)(1)(A)(ii). Section 512(c)(1) was not at issue in Io because Io never actually even sent a takedown notice; after spotting infringing files on Veoh's site, it decided to proceed directly to an infringement complaint. (Is this another case of, perhaps, some pre-1998 law school grads being a menace?) Section 512(c)(1) says that, in order to keep its immunity, an ISP hosting infringing content must do a few things, one of which is "act[ ] expeditiously to remove, or disable access to, [infringing] material" if it becomes "aware of facts or circumstances from which infringing activity is apparent." Congress called this a "red flag" provision in the legislative history; an ISP cannot simply ignore red flags indicating infringement and wait for the takedown notices to flow in, if it wants to preserve its immunity.

Does choosing not to use a filter that would screen out additional copies of a work cited in a takedown notice constitute ignoring a red flag? Io suggests that the answer may be yes, and if so, YouTube may be in trouble.

Posted by Bruce Boyden on August 28, 2008 at 06:57 PM in Information and Technology, Intellectual Property | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Io v. Veoh: Harmful to YouTube?:


Michael, I had a line in my post at one point about how the vicarious liability stuff would appeal only to copyright geeks, but I deleted it. I see that you fell right into my trap (muhahahahaha!).

More seriously, the word "somehow" in my phrasing of the test that "YouTube's ability to sell ads was *somehow* enhanced by the infringing content" was meant to weasel out of a discussion of what the standard exactly is, because as C.E. Petit points out, it's not exactly clear in the 9th Circuit, and even less clear in other circuits that haven't had many recent cases on this (e.g., the 2nd, where the YouTube case is). I don't think even post-Fonovisa the 9th Circuit's test is quite "more content = more ad revenue" though, because there needs to be some connection between the infringing content and the increased ad revenue (or revenue from any source). The Fonovisa court drew a sort of common-sense conclusion that the attendee fees in that case were increased by the availability of infringing records, but in Ellison, as C.E. points out, the 9th Circuit got a little more rigorous about the need for proof. "[T]here is no evidence that indicates that AOL customers either subscribed because of the available infringing material or canceled subscriptions because it was no longer available." That sounds like it's almost calling for trademark-litigation-style surveys. Yuck!

C.E., I'm hesitant to say that ignoring red flags eliminates s 512(i) qualification for the safe harbor, for two reasons. One, doing so threatens to make s 512(c)(1)(A)(ii) and (iii)--the red flag provisions--superfluous. Two, the policy that must be reasonably implemented under 512(i) is one that provides for termination of the accounts of repeat *infringers*. The purpose of filtering is to remove the same content, rather than the same users. If there's an obligation not to partially filter, I think it applies to ISPs who have received a takedown notice, or perhaps multiple takedown notices for the same content, because the notice gives them awareness of facts indicating infringement under 512(c) and (d).

Posted by: Bruce Boyden | Aug 29, 2008 1:04:17 PM

Mr Risch, your comment points out a partially unresolved problem with the extent of Fonovisa. I made the same argument in Ellison, which the 9th Circuit rejected (we won on other grounds). However, a fair reading of the Supreme Court's opinion in Grokster agrees with you, if not so loudly that the remand managed to avoid the issue.

Prof Boyden's initial post also implicates the very (unsound) structure of § 512. Assume, for the moment, that the withholding of filtering technology on a holder-specific bases (no license, no filter) somehow violates § 512. But what part of § 512 does it violate? The only place I can see that it does so is § 512(i), going to the reasonableness of the service provider's policies (and implementation of those policies). In some ways, this makes sense; it's akin to an unclean-hands objection. The problem is that failure of any part of § 512(i) wipes out all of the service provider's entitlement to DMCA safe harbors, even irrelevant ones, under Ellison's reasoning. As someone who represents copyright holders, I think this is probably the right result... but it's certainly not the result that most people think would be right, particularly for service providers that have multiple services.

Posted by: C.E. Petit | Aug 29, 2008 11:35:07 AM

Question on this: "The second prong would still need to be demonstrated, and that would likely require a showing that, e.g., YouTube's ability to sell ads was somehow enhanced by the infringing content. Maybe that would succeed, maybe it wouldn't."

Why must the showing be so direct? Can't it just be that more content = more ad revenue? Or is that what you are saying? Cases like Fonovisa seem to allow a pretty loose financial gain showing (there it was trade show booth revenue if I recall).

Posted by: Michael Risch | Aug 29, 2008 9:14:19 AM

The comments to this entry are closed.