In May 2020, West Publishing and Thomson Reuters filed a copyright action against ROSS Intelligence LLC, alleging that ROSS (through a third party) scraped content from WestLaw to start its own Artificial-Intelligence-based legal research platform. (ROSS has since ceased operations but persists solely to litigate this case.)
Down the line, I imagine the case may raise some interesting questions about AI and copyright. For example, what are the copyright implications of ROSS's use of Westlaw's copyrighted compilation of otherwise public domain materials to train an AI? Isn't that fair use (talk about transformative!)? If not, what are the damages? And so on.
For now, ROSS has moved to dismiss on the ground that West failed to state a claim of infringement, arguing among other things that the attorney-written headnotes in Westlaw are not creative and are not entitled to copyright (good luck!) and that the complaint failed to specify exactly what copyrighted material was copied.
Chief Judge Stark Requires Copyright Contentions
At oral argument, while addressing whether West had sufficiently identified its copyrighted material, Chief Judge Stark stated that the parties would be required to build into the schedule time for copyright infringement contentions, just like in a patent action:
THE COURT: . . . In patent cases, we usually have infringement contentions at a fairly early stage in the discovery process. Isn't that part of the answer here? [Plaintiffs] will be required, if they survive this motion to dismiss, to, in a fairly expeditious manner that we would build into the schedule, serve something like infringement contentions on you that make clear, here is what we contend is protectable about, you know, our intellectual property, and here is what we contend [FOSS] copied.
I took a look at a number of recent copyright cases, and I didn't see provisions for copyright infringement contentions (including in Chief Judge Stark cases). So this may be a new procedure.
Normally contentions in non-patent actions are developed through contention interrogatories, just as they were in patent actions up until the District of Delaware issued its Default Standard for Discovery in 2011. (The Default Standard set forth the practice of initial and final invalidity contentions in patent cases.)
I'm curious about what those copyright contentions may end up looking like. It seems to me that copyright contentions might actually be more analogous to trade secret identifications than to patent contentions, given how many different pieces of copyrighted content are likely involved.