Electronic Arts, Inc. (“EA”), owner of the $4 billion John Madden Football videogame franchise, thought it had a pretty good defense when Robin Antonick filed suit in the Federal Court for the Northern District of California, claiming that EA had infringed his computer software copyright.
Specifically, Antonick didn’t appear to have a copy of the copyrighted source code, i.e. his original program, so the jury simply could not conduct a side-by-side comparison of the allegedly infringed and allegedly infringing works. EA felt this would make the substantial similarity test an impossible hurdle for Antonick to overcome. But the jury didn’t see it that way. Will District Court Judge Charles Breyer let the jury’s verdict stand?
In 1983, Antonick, a former college football player and computer programmer, approached a then-fledgling EA with a new way to create football videogames. Antonick used sophisticated models of player behavior to simulate more realistic gameplay experience. The result (presumably after a few turduckens and a licensing deal with former NFL coach John Madden) was EA’s 1988 launch of John Madden Football, which became a wildly successful product for the Apple II and other early personal computing platforms.
Antonick’s compensation from EA included royalties on future versions of the software that were based on his original work. According to the complaint, in 1990 EA sought to avoid some of these royalties by creating a separate version of the game for the new Sega Genesis platform. EA assured Antonick that his intellectual property was not used to create the new software, but Antonick alleged that, contrary to those assurances, EA copied his original Apple II source code.
According to Antonick, he only learned of this illegal copying in 2009, as a result of some careless public statements by an EA executive. In 2011, Antonick filed suit against EA for breach of contract and fraud. Although the complaint did not include a copyright infringement count, the parties and the judge appear to have agreed that the contract claim turned on copyright law principles. But in bringing what was essentially a copyright claim, Antonick had one major problem. He could no longer locate his original copyrighted source code. So how could he prove that it was copied?
Locating the Source
The plaintiff’s inability to locate original source code is not an unprecedented problem in the context of computer software copyright infringement. The same thing happened in the First Circuit case of Airframe Systems, I v. L3 Communications , 685 F.3d 100 (1st Cir. 2011). The plaintiff produced aircraft maintenance software in the 1980’s and subsequently registered it with the copyright office. However, by the time of trial, the only iteration of the source code still available was the updated and unregistered 2009 version. The First Circuit affirmed summary judgment for the defendant because, without the original source code that had supposedly been accessed and copied by the defendant, the finder of fact could not conduct a side by side comparison between the allegedly infringed source code and the allegedly infringing source code to see if they were substantially similar.
Antonick’s End Around
Unable to present the jury with a side by side comparison of his original source code and EA’s Sega Genesis Source code, Antonick tackled his burden of proof in a different manner. His expert took some surviving original binary code, the machine-readable ones and zeros that were compiled from his source code, and used that data to create a visual representation of the kind of football play formations the original Apple II program would create. He then compared that visual representation to a visual representation of the play formations that EA’s Sega Genesis source code would create. Thus, according to EA, Antonick invited the jury to work backwards from the visuals and infer that EA must have copied the absent source code because the visual outputs of those codes looked alike. The jury apparently accepted this invitation and, on July 23, 2013, found in favor of Antonick because there were “substantial similarities . . . in the source code.”
Was the jury confused?
EA was not only unhappy with the verdict, but perplexed. How could a jury possibly find that two source codes were substantially similar when that jury neither saw nor compared those source codes? On August 15, 2013, EA filed a Motion for Judgment as a Matter of Law (aka Judgment notwithstanding the verdict), which asks the court to ignore the jury’s verdict, and which argues, inter alia, that the jury was hopelessly confused by the presentation of impermissible evidence in at least two related ways.
First, by permitting Antonick to prove his case with the visual representations of play formations, the court blurred the line between a computer program’s underlying source code and its visual output. The visual output of a computer program can be copyrighted, but only separately from the program as a distinct audio-visual work. This separate registration makes sense, because what we see on the screen is often directly related only to a very small part of the underlying programming, and also because two very different programs could in theory create similar visual images. Thus, some courts have opined that the use of visual output evidence to prove the contents of source code is dangerously prejudicial, because a jury faced with highly technical evidence is likely to gravitate towards the simplicity of an image, however tangentially relevant.
Second, EA claims that Antonick’s presentation improperly encouraged the jury to find copyright infringement by comparing uncopyrightable elements. EA claims that the visual displays of the play formations were not copyrighted (because they were prepared by the expert) or copyrightable (because they are common football formations). Thus, the jury inferred copyright infringement from its comparison of uncopyrightable materials.
But the issue of uncopyrightable elements really goes deeper than that. Courts have long held that certain aspects of a source code are not copyrightable, for example portions of the code that are dictated by design standards, or that are required by technical limitations, or that are simply in the public domain. Typically, these elements are identified for the jury and “filtered” from the analysis. However, when a visual display is used to prove the contents of the source code, and the source code is not in evidence, it is possible that what the jury sees was created not by original copyrightable programming, but by portions of the code that could not be copyrighted in the first place.
In opposition to EA’s motion, Antonick argues that EA’s complaints are much ado about nothing. According to Antonick, the jury did what it was supposed to do in the computer software context: it relied on expert testimony to compare the surviving Apple II code and the Segal Genesis source code, thus making the side by side comparison that is required for a finding of copyright infringement. According to Antonick, “EA’s statements to the contrary represent nothing more than its attorneys’ failure to understand key technical aspects of the case.”
Briefing on the motion was completed on September 6, 2013, and Judge Breyer has the matter under advisement.