Google Defeats Copyright Lawsuit Over Waze Data (Forbes Cross-Post)
The basic copyright rule is clear: facts are not copyrightable; factual compilations can be. However, this simple rule masks considerable nuance. What is a “fact,” how does it differ from “non-facts,” what does it mean to “compile” facts, and when is a compilation sufficiently original to become copyrightable? These questions are more epistemological than legal, so not surprisingly, the associated legal disputes routinely baffle judges. As a result, the copyright caselaw regarding facts and compilations is confused–and confusing.
These issues surfaced again in a recent case where Google defeated a copyright challenge over data used in its Waze navigation application (Google bought Waze in 2013). The plaintiff, PhantomALERT, offers a GPS-based navigational app that competes with Waze. Both apps use databases containing “the location of traffic conditions, road hazards, and traffic enforcement monitors, such as speed cameras” (what the court calls a “points of interest database”). PhantomALERT alleged that Waze ripped off its points of interest database, as evidenced by the alleged presence of fake points of interest created by PhantomALERT appearing in Waze’s database.
This fact pattern resembles Feist v. Rural Telephone Service, the seminal 1991 Supreme Court opinion involving the copying of telephone “white pages” data. In that case, a phone book publisher copied a rival’s white pages data, including fake listings, to produce a competing phone book. The Supreme Court, in an opinion I’ve routinely characterized as the single most important copyright opinion ever, held that individual white pages listings were just facts and never copyrightable. The court further held that compilations of white pages listings could be copyrightable if they showed sufficient originality, but presenting white pages listings in alphabetical order by surname wasn’t original at all.
The Feist case casts a long shadow on the PhantomALERT case. The court held that individual points of interest were facts and therefore never copyrightable. PhantomALERT argued that it exercised some judgment deciding where to place each point of interest on its map and how much advance notice to give drivers about each point of interest. The court says the location decision is driven by functional considerations, which I infer means that PhantomALERT sought to be as factually precise as possible to improve the app’s functionality. The court also says there’s no evidence Waze copied any of PhantomALERT’s judgments about where to locate the points of interest or how much notice to give drivers.
But what about PhantomALERT’s overall compilation of points of interest? Per basic copyright law, PhantomALERT ought to have a compilation copyright for its database as a whole. The judgments PhantomALERT made to prepare a detailed map surely are significantly more extensive than the simplistic alphabetization of white pages info. However, the compilation copyright would be “thin” in the sense that it would only prevent wholesale verbatim copying. Any other implementation shouldn’t be copyright infringement because it doesn’t copy PhantomALERT’s original contributions.
Applying these basic principles, the court says there doesn’t appear to be any originality in how PhantomALERT organized the points of interest database, but PhantomALERT may have exercised enough judgment selecting which points of interest to include in the database. As evidence of PhantomALERT’s editorial judgment about selecting or excluding facts, the court gave the example of how PhantomALERT may delete speed traps from its database if it believes those traps don’t pose a significant risk to drivers.
This copyrightable interest in selecting factual data still doesn’t help PhantomALERT. The court says:
the issue is whether the defendant copied the element of the compilation that was original. The creativity that may have been reflected in PhantomALERT’s selection of Points of Interest is not implicated by merely incorporating the “raw data” into Defendants’ databases along with other data already contained in those databases.
I read this to say that even if Waze took 100% of PhantomALERT’s factual selections, there’s no copyright infringement if Waze adds its own selections to the mix. As a result, the court dismissed PhantomALERT’s case but gave PhantomALERT a chance to try again on this point.
We’ll have to see how PhantomALERT’s second attempt fares before drawing conclusive lessons from this case. If PhantomALERT gets better traction next time, it could cause a lot of ripple effects through the transportation data ecosystem. In the more likely scenario where this case fails again, it will provide more comfort to transportation data providers that they aren’t stepping on unexpected copyright landmines as they remix rivals’ “facts.”
Case citation: PhantomALERT, Inc. v. Google, Inc., 3:15-cv-03986-JCS (N.D. Cal. Dec. 14, 2015)