Automated screening will remove entire categories of goods when the description is trademarked

Back in 2010, Tiffany v. eBay held that online marketplaces were not liable for trademark infringement, false advertising, or trademark dilution when a third-party seller sold counterfeit goods. SHOP SAFE would essentially overturn the Tiffany decision and open up third-party sellers to liability. Trademark owners would now be able to pursue both the infringing party, the third-party seller, and the online marketplace. The bill does not require that trademark holders send takedown notices for these situations.

Well if third-party sellers aren’t selling counterfeit goods then they should be fine, right? Not necessarily. SHOP SAFE would require online marketplaces to implement systems or “reasonable proactive technical measures” that pre-screen for trademark infringement. These technological measures would likely involve a type of automated screening or filtering, which is susceptible to error and overcorrection and products that are in compliance could unintentionally be flagged. These pre-screen measures have already not been very effective in the copyright space in distinguishing between infringing and non-infringing uses, and we don’t see why that would improve with trademark.

One of the biggest issues with SHOP SAFE is it’s circular definition for what is considered “counterfeit” or a “counterfeit mark”. The bill defines it as “a mark that is counterfeit”. If we look to the Lanham Act, there is no definition. We can find a definition in the United States Criminal Code 18 U.S.C. 2320(f)(1), but SHOP SAFE does not amend or even point to the criminal code. We are likely left to look at common law and counterfeit can be interpreted broadly enough to include “knock-offs”, items that have a variation of the mark, or whatever other definition the courts seek to apply.

Additionally, there are a lot of words that sellers use to describe goods that are actually trademarked. For example, Sealed Air Corporation owns the trademark for bubble wrap, Gerber for onesie, Wham-O for frisbee, Mathomos for lava lamp, and the list goes on and on. If a seller uses any of these words in the product name, description, a tag, etc. the automated filters would likely catch it and that may be enough to remove the entire listing. If a seller is selling bubble wrap they would have to use a different, non-trademarked name. But now, how would buyers find what they want? If you search “bubble wrap” on an online marketplace, listings for bubble wrap used in packaging come up. If you search for “bubble pack”, a synonym, you will probably find packs of liquid bubbles, miscellaneous bubble toys, and maybe some bubble wrap. Perhaps a more frustrating issue is if a seller is attempting to list something like “apple mug” , a coffee mug shaped like an apple. The filters then catch Apple, like Apple Inc., and take down the listing. Whether sellers are creating their own products or reselling used products, they would be affected. Buyers would suffer from a normal search becoming immensely complicated and confusing.