Yesterday I was fortunate to speak about the complexities of trade dress laws at INTA’s 3D printing conference, and how unlikely it is that an average user could be expected to know whether they are violating them when they create and share copies of a product using a 3D printer or scanner. Most of my talk was about how hard it is to navigate questions of functionality and distinctiveness that arise in product design and product packaging. But I also briefly touched on the fact that there is no DMCA (or Canadian Copyright Modernization Act) notice equivalent for trademarks in the US or Canada, and I suggested the lack of a legislated safe harbor for ISPs could be cause for concern.
After my talk, I was asked whether the 2010 Second Circuit Court of Appeals decision Tiffany Inc. v. eBay Inc. effectively addresses the concern about a lack of safe harbor for 3D ISPs – specifically sites that allow for the creation, uploading and sharing of 3D content. (And I admit I didn’t even remember the case when I was asked about it. I personally blame the brain atrophy that comes with being an in-house lawyer.) Later I had a forehead-slap moment, so I thought I’d take the opportunity to discuss it here, and explain why I think we need an ISP safe harbor for contributory trademark liability.
In Tiffany v. eBay, eBay was sued over the sale of counterfeit Tiffany merchandise on its auction website. eBay had (and has) a system in place called VeRO (Verified Rights Owner) where rightsholders could submit notices of infringement when they saw such activity taking place. But Tiffany argued that it wasn’t enough – Tiffany claimed eBay knew that counterfeiting was rampant, and was both directly and secondarily liable for the counterfeit products being sold through its platform. I won’t go deep into the reasoning behind a 5 year old case (there are plenty of summaries out there), but there were a few big points the case turned on. Most importantly, although eBay had ‘general knowledge’ of infringement occurring on its site, it didn’t have ‘actual knowledge’ of infringement and it took ‘appropriate steps’ when it was notified of alleged infringement by quickly removing the content:
For contributory infringement liability to lie, a service provider must have more than a general knowledge or reason to know that its service is being used to sell counterfeit goods. Some contemporary knowledge of which particular listings are infringing or will infringe in the future is necessary.
Don’t get me wrong – that is a massive comfort to ISPs offering 3D services (like my company Matter and Form’s forthcoming Cashew 3D presentation platform). So do we still need something like the DMCA safe harbor for trademarks to protect 3D content ISPs? Yes, I think there’s major reasons to still want one:
1) eBay had built a complex ‘fraud engine’ and was spending $20m per year on its program to fight counterfeiting, a fact which was noted by the Court. The DMCA, by contrast, creates a set of fairly simple steps that a responsible company can follow in order to be protected by its copyright safe harbor. eBay’s expensive anti-counterfeiting efforts clearly helped it secure a positive verdict – the Court spends a lot of time on it in the decision. But the opposite was true for the Alibaba-like site Tradekey which, in the wake of Tiffany v. eBay was successfully sued by six luxury brands for contributory infringement despite never having received a formal complaint from the brands about any offending products. The decision was largely based on finding that the company was actively promoting counterfeiting by its customers. For all the ISPs who lie somewhere in the middle, neither promoting counterfeiting, nor spending millions on an anti-counterfeiting program, the Tiffany decision still leaves a few questions. In a later California appellate decision, Tre Milano v. Amazon.com, a very different system of dealing with counterfeiting notices employed by Amazon was also upheld. Amazon’s system meant that alleged counterfeit listings sometimes remained up for weeks after a notice was issued, and many notices were rejected for lack of evidence. But Amazon’s system was also very sophisticated, involving over 100 employees and an investigation of each case. And the appellate court in Tre Milano v. Amazon.com again spends considerable time talking about Amazon’s extensive anti-counterfeiting measures. I’d also argue that none of these previous cases are entirely helpful when considering alleged trade dress infringement issues, since these cases of counterfeiting are more clear-cut than what we have to deal with in the 3D model world. What system do you use when the existence of a legitimate claim is entirely possible, but so uncertain? And that brings me to the next point…
2) The DMCA actually has built-in balancing mechanisms in the notice and takedown regime that at least attempt to protect end users, and guard them against spurious claims. A user who believes their content has been wrongly removed can file a counter-notice. Upon receipt of such notice, their identity is revealed to the party who filed the original notice and that party has 10 days to inform the ISP that they have filed an action in court seeking to ‘restrain’ the allegedly infringing party from putting the offending material online. But how does this sit with Tiffany v. eBay? Tiffany said that eBay’s removal of allegedly offending content was sufficient action upon receiving notice. It said nothing about allowing counter-notices. But if an ISP allows for counter-notices for alleged trademark infringement, without the DMCA’s statutory 10 day requirement for a rightsholder to either initiate an action or back down, putting material back online could once again place the ISP at risk of secondary liability (since now the ISP has ‘actual’ knowledge). Given how tricky it is to establish whether or not a valid trade dress right might exist in a physical object design in the first place, the Tiffany decision could leave 3D ISPs in the uncomfortable position of siding with the party filing a complaint simply to avoid the possibility of liability.
The Tiffany decision also leaves an interesting hole when applied in our 3D modeling world – one that I personally don’t have that much of a problem with. The Court found eBay was not directly liable, noting that eBay would have had a nominative fair use right to use the Tiffany name for at least genuine Tiffany merchandise, and it promptly removed allegedly infringing listings upon receiving a complaint. But there was little question in that case that there was direct infringement occurring by the eBay users selling counterfeit products. Tiffany had hired a company to purchase supposed Tiffany products posted on eBay and was able to prove massive amounts of counterfeiting was occurring. On the other hand, aside from the difficulty in establishing trade dress rights in product design, uploading a 3D model for free viewing/downloading on a site like Cashew or Thingiverse might be considered noncommercial use by the user, in which case no direct infringement would actually be occurring. As a result, barring some other argument, no contributory infringement would be occurring either.
What if the platforms could not be held liable for the actions of their users? I believe Cory Doctorow defends something like that.