The Progress & Freedom Foundation Blog

This is true even though--for journalists and wonks--the interesting part of a judicial decision is usually its interpretation of the law. Cases almost always have materially differing facts, so the court's interpretation of the law is usually the most broadly relevant aspect of the decision. But before turning to Veoh's interpretation of the law, a critical question must first be answered: Did the opinion granting summary judgment correctly identify the "undisputed" facts before applying the law to them?

I put the term "undisputed" in quotations because it characterizes a more complex reality. Summary judgment is a critical component of the federal judicial process. Nevertheless, federal courts cannot grant summary-judgment when doing so would usurp the fact-finding role of the jury. Summary judgment would become an unconstitutional usurpation of power were it just a way for judges to rule in favor of the party they would have voted for, had they been federal jurors, rather than federal judges.

To avoid encroaching on the jury's fact-finding powers, judges grant summary only when all material facts in a case are "undisputed" and the application of the law to those facts dictates a decision for a party on some or all claims or defenses in a case. But the few cases in which there really are no disputes about any facts material to any case, claim or defense cannot explain the prevalence of summary judgment.

Nevertheless, courts can grant summary judgment even in cases in which conflicting or ambiguous evidence creates actual disputes about what the facts are or what inferences should be drawn from them. Such disputes can be made immaterial if all reasonable disputes about what the facts are, and all reasonable disputes about what inferences should be drawn from them are resolved in favor of the non-moving party. Consequently, as Judge Matz put it in Veoh, "[I]n ruling on a motion for summary judgment, the nonmoving party's evidence 'is to be believed, and all justifiable inferences are to be drawn in [that party's] favor." Judges granting summary judgment must thus be reversed on reconsideration or appeal if they fail to correctly describe the set of facts that must be deemed "undisputed" when deciding a summary-judgment motion.

Policy analysts can usually skip the step of cross-checking an opinion's account of the "undisputed" facts of the case. Ordinarily, that task will be extremely difficult: it will mean carefully reviewing a novel-length set of briefs supporting cross-motions for summary judgment and conflicting Proposed Statements of Facts and Conclusions of Law. It may even be impossible: a confidentiality order may ensure that only counsel to parties to the case, the court, and the appellate court(s) will be able to cross-check a summary-judgment ruling's account of the "undisputed" facts of the case.

But in the opinion granting summary judgment on the Veoh defendant's eligibility for the DMCA's "hosting-site" safe harbor, a serious problem with the Court's interpretation of the "undisputed" facts seems evident in the opinion itself. This rarely happens, but when it does, it can not only signal reversible error, but also other defects in the characterization of other facts lurking beneath the surface.

A summary of the relevant legal issues can clarify the problem with Veoh's account of the "undisputed" facts. The question before the court in Veoh was whether Veoh was entitled to the liability-limitations granted by the DMCA's Â§ 512(c) "safe harbor" for sites that host third-party content. While Â§ 512(c) can protect hosting sites that comply with certain basic requirements, (like responding to take-down notices from copyright owners), it was not meant to let hosting sites sit in a "safe harbor" while copyright owners play an extended game of takedown-notice-whack-a-mole with their infringing users that either undermines the vitality of their federal rights or forces them to sue individual Internet users--again.

To foreclose such results, Congress added two limitations to the hosting safe harbor. The "culpability limitation" in Â§ 512(c)(1)(A) (sometimes called the "red-flag" test), denies a "harbor" to hosting sites that "have actual knowledge that the material or an activity [on their site] is infringing" or lack such knowledge, but become "aware of facts or circumstances from which infringing activity is apparent." The "control limitation" in Â§ 512(c)(1)(B) denies a "harbor" to any hosting site that "receive[s] a financial benefit directly attributable to the infringing activity, in a case in which the service provider has the right and ability to control such activity...."

A critical question in Veoh was thus whether Veoh had triggered the culpability or control limitations on the hosting safe harbor. Any facts or inferences relevant to culpability or control were thus highly material and--if disputed--would have had to have been given the reasonable interpretation most favorable to UMG. On its face, the Veoh opinion seems to fail to apply this familiar standard when interpreting the "undisputed" facts relevant to the most powerful evidence discussed in the opinion.

UMG showed that Veoh had paid search engines to display sponsored links that would direct to Veoh users who had searched for specific works owned by UMG, (like the video for artist 50 Cent's Grammy-nominated song "Candy Shop")--works that were not legally available on Veoh's site, but may have been made available there illegally by Veoh users. Here is the unanimous Supreme Court in Grokster, explaining why evidence like this might be dispositive of an intent-based inducement claim--and thus highly relevant to the knowledge-based or control-based limitations on the Â§ 512(c) safe harbor:

Evidence of "active steps taken to encourage direct infringement,"... such as advertising an infringing use... show an affirmative intent that the product be used to infringe....

Here, (with citations omitted), is the Veoh opinion's attempt to characterize the implications of Veoh's conduct for summary judgment:

UMG next points out that Veoh paid search engines to have a link to its website appear in search results when users used certain search terms, including the names of specific UMG-controlled music videos. Veoh, however, presented rebuttal evidence that the five artists referred to in the search terms UMG identified are Sony BMG artists whose videos Veoh streams with SonyBMG's consent.
At the hearing, UMG stressed that one of the keywords that Veoh purchased was "50 cent candy shop music video," and asserted that UMG controls the copyright to that video. UMG does not, however, dispute the evidence that SonyBMG controls the rights to at least some of 50 Cent's works, and UMG acknowledges that SonyBMG hosts two 50 Cent videos.

UMG misconstrues the deposition testimony of Veoh's former head of marketing Jennifer Betka, by stating that she admitted that Veoh would only buy such terms if it knew that the infringing videos were present on its site. In fact, Betka stated she was speaking in general terms, was not "close to" the purchasing of the search terms by Veoh, and was very unfamiliar with this level of detail against Veoh's search engine media habits...." Furthermore, the keywords were associated with links to the authorized SonyBMG videos.

On its face, the court's analysis of these facts--and the inferences that can be reasonably drawn from them-- seems to violate basic rules of summary judgment. Simply put, the Court seems to resolve disputed facts and inferences in favor of the moving party.

No matter how they are interpreted, these facts seem to put Veoh in a bad light. Veoh seems to have been playing bait-and-switch. But the critical question is: bait and switch to what? Two sets of inferences seem plausible:

The pro-Veoh inferences that could be reasonably drawn would be those proposed by the Court: Veoh was paying to lure to its site consumers who might realize that they had been duped, and then stay and do business with the site that tricked them.

The pro-UMG inferences that could be reasonably drawn would be these: Veoh was paying to lure to its site consumers who would think that the sponsored link was misdirected (or the video moved), search Veoh, find an unauthorized "50 cent candy shop music video" and forget about the initial "misdirected link."

Both sets of inferences seem plausible, and neither seems foreclosed by any evidence cited in the opinion. Granted, the first set of inferences seems very strained and suspect, and even if the Betka testimony does not foreclose it, even a "general" practice of buying sponsored links only to content actually available on Veoh would enhance the plausibility of the second set of inferences.

On summary judgment, I don't see how a court could adopt the pro-Veoh inferences that would direct any chicanery toward consumers, rather than UMG. By contrast, were the UMG-favoring inferences presumed, then this and similar episodes would seem highly relevant to intent-based inducement liability, the culpability-based harbor limitation, and the control-based harbor limitation. (You can certainly control whether you direct to your site persons looking for content that you may have, but not legally.) Consequently, this seems like a pretty serious, blatant and dispositive error--particularly since failure to apply the relevant summary-judgment standards to the facts of the case is an error of law reviewable de novo on appeal.

As for the legal analysis in Veoh, I think that the factual problem just discussed severely undermines its credibility. Moreover, Ben and Chris have noted many of the problems with it. I would thus note only one more point--not a problem, but a it's-kind-of-silverish lining for those concerned about the long-term implications of a decision like Veoh.

It would be very problematic--and contrary to Congressional intent--were Â§ 512(c) interpreted to harbor hosting sites that insist on re-running the early YouTube's takedown-only experiment to re-confirm that it still generates massive infringing use that must later be remediated by other means. And while it is not clear to me from the Veoh opinion whether Veoh could fairly be said to present such a case, such concerns are quite valid and they are heightened by some of the dicta in Veoh.

Fortunately, any permissible interpretation of the culpability limitation on the Â§ 512(c) harbor should grow more powerful over time. This is a critical point that the anti-copyright crowd tends to forget: no criminal would ever be convinced unless judges and juries were both empowered and able to infer unspoken culpability from facts and circumstances.

For example, the law presumes that people intend the obvious consequences of their own actions. If so, then neither Â§ 512(c) nor Â§ 512(d) can long "harbor" an endless procession of hosting sites or file-sharing-program distributors that insist on re-running (with minor variations) the early-YouTube or Napster/Aimster/Grokster experiments. Legal rules could be questionable if they required people to make detailed predictions about the distant future. But not if they required people to learn from the past. Consequently, the day should come, if it has not already, when yet another attempt to re-run Napster or Morpheus should--in itself--suffice to waive the red flag, drain the harbor, and support inducement liability under Grokster.

So in the long run--and regardless of whether it is sustained or overturned--the Veoh decision's most important contribution may be to memorialize that takedown-plus-hash-based-filtering is an ineffective strategy that should never again be pursued by hosting sites dedicated to building businesses based upon lawful, non-infringing uses of their services. Let that be a lesson to those who follow.