http://ipkitten.blogspot.com/2025/04/update-reuters-v-ross-trial-postponed.html
The IPKat recently covered (see here) the decision of the Delaware District Court (the Court) in Reuters v ROSS (Case 1:20-cv-00613, ECF 770), in which it held that a fair use defence was unsuccessful on a summary judgment motion for copyright infringement. The trial, which was due to take place next month, has now been postponed pending an interlocutory appeal.
Background
By way of reminder, ROSS Intelligence Inc (ROSS), a legal research platform, needed a database of legal questions and answers to train its AI (but not generative AI) search tool, and initially sought a licence from Westlaw, but Reuters refused.
Ross therefore entered into an agreement with a company called LegalEase to provide training data in the form of “Bulk Memos”, which were effectively compilations of legal questions prepared by a team of lawyers using Westlaw headnotes. This was found not to constitute fair use, which could have potentially very important ramifications for the other AI/GAI-related infringement cases before the US courts.
Appeal
Last week, the Court granted ROSS’ motion for an interlocutory appeal of its decision, meaning that this will be the first case to reach an appellate court (here, the Third Circuit) on the fair use/training issue (albeit in respect of “simple” AI training).
In its Order, the Court recognised that there were substantial grounds for difference of opinion on controlling legal issues in the case, and certified the following two questions to the Third Circuit:
- whether the Westlaw headnotes and numbering system are original; and
- whether Ross’s use of the headnotes was fair use.
However, pursuant to 28 U.S.C. § 1292(b), the Third Circuit also needs to agree to hear the appeal, and ROSS filed its petition accordingly this week.
In summary, ROSS argues that the Court’s ruling is doctrinally flawed regarding their application of the fair use test, because it failed to answer the central question of the first of the four fair-use factors, which looks at the purpose and character of the use in question.
IPKat readers will remember that the Court found it impossible that ROSS could engage in transformative intermediate copying because it was developing a product which competed with the rightsholder’s. ROSS contends that it failed to address the question of whether the product added something new with a different character than Reuters’ manual headnote sorting system.
ROSS also argues that the Court erred in respect of the fourth factor, calling its determination that “the public has no right to Thomson Reuters’s parsing of the law” to be “jaw dropping”.
Finally, ROSS challenges the Court’s determination that the first and fourth factors were the ones to be given the most weight in the fair use analysis, and suggests that greater weight should have been placed on its de minimis use of the headnotes (it only used 0.076% of the total headnotes) as part of the third fair-use factor.
Public benefit
Arguably the most interesting parts of ROSS’ petition are those which do not concern strictly legal arguments. In particular, ROSS argues that permitting the interlocutory appeal would serve the public interest, citing arguments by OpenAI (which is not a party to this case) that a failure to hold that AI training is fair use will harm “national security” and mean that the “race for AI is effectively over”, leaving “foreign corporations” (possibly a reference to DeepSeek) the winners.
Comment
This is clearly a case to watch for all copyright practitioners (and those interested in the interplay between IP law and AI generally). It does seem that things seem to be coming to a head soon, as just last week Jack Dorsey, a co-founder of Twitter (now X) posted a now deleted tweet which just stated “delete all IP law”. Elon Musk promptly agreed.
Between this case, the trial of Getty v Stability AI in the UK (see IPKat here) which is due to be heard in June this year, the imminent publication of the EU Commission’s final draft of the GPAI Code of Practice, and Part 3 of the US Copyright Office’s report, which will focus on AI training, it is likely to be a very informative few months in this space.
Content reproduced from The IPKat as permitted under the Creative Commons Licence (UK).