AI Copyright Infringement: Early Ruling Says Yes
- Examining the implications of a key AI-copyright decision for AI providers and the future of copyright law.
- Artificial intelligence (AI) providers are facing numerous lawsuits from copyright owners who allege that AI models infringe their copyrights when trained using copyrighted works. In a significant development,...
- AI-copyright decision to address whether the fair use defense shields an AI model provider from copyright infringement claims.
AI Copyright Litigation: Thomson Reuters v. Ross and the Fair Use Defense
Table of Contents
- AI Copyright Litigation: Thomson Reuters v. Ross and the Fair Use Defense
- AI Copyright Litigation: Thomson Reuters v. ross and the Fair Use Defense – Q&A Guide
- Introduction
- General Questions
- Case Details
- Q4: What are the facts of the Thomson Reuters v. Ross case?
- Q5: What did Thomson Reuters allege in their complaint against Ross Intelligence?
- Q6: How did Ross Intelligence obtain the data used to train its AI model?
- Q7: What was Ross Intelligence’s argument in defense against the copyright infringement claim?
- Fair Use Analysis
- Q8: What are the four factors considered in a fair use analysis?
- Q9: How did the court analyze the “purpose and character of the use” factor in Thomson Reuters v. Ross?
- Q10: How did the court analyze the “nature of the copyrighted work” factor?
- Q11: How did the court analyze the ”amount and substantiality of the portion used” factor?
- Q12: How did the court analyze the “impact on the copyrighted work’s value or potential market” factor?
- Q13: Why did the fair use defense fail in the Thomson Reuters v. Ross case?
- Implications for AI and Copyright
- Q14: What implications does the Thomson Reuters v. Ross case have for AI providers?
- Q15: How might this case effect future AI copyright cases?
- Q16: Will the outcome always be the same in other AI-copyright cases?
- Q17: How does the court’s reasoning differentiate between generative and non-generative AI?
- Q18: What business takeaways can be derived from the Thomson Reuters v. Ross case?
- Summary Table: Fair Use Factors in Thomson reuters v.Ross
- Conclusion
Examining the implications of a key AI-copyright decision for AI providers and the future of copyright law.
Artificial intelligence (AI) providers are facing numerous lawsuits from copyright owners who allege that AI models infringe their copyrights when trained using copyrighted works. In a significant development, the U.S. District Court for the District of Delaware ruled in favor of a copyright holder in Thomson Reuters Enterprise Center GmbH et al v. ROSS Intelligence inc. (Thomson Reuters v. Ross). This decision casts doubt on whether the fair use defense will protect AI providers from liability.
The Thomson Reuters v. Ross case marks the first major U.S. AI-copyright decision to address whether the fair use defense shields an AI model provider from copyright infringement claims. The court rejected Ross IntelligenceS fair use defense, asserting that, based on the specific facts, using copyrighted data to train its AI model constituted direct copyright infringement and was not fair use. While the outcome may differ in other AI-copyright cases, it establishes a foundational precedent in copyright jurisprudence concerning AI and fair use.
Facts of the Case
In May 2020, Thomson Reuters, the owner of Westlaw, filed a complaint against Ross intelligence, alleging intentional copyright infringement. Ross had developed a legal research search engine intended to compete with Westlaw.Unlike generative AI tools,Ross’s tool was an AI search engine that answered legal questions by providing relevant judicial opinions. To train its AI tool, Ross aimed to use westlaw’s headnotes and numbering system as a database of legal questions and answers. Thomson Reuters owns copyrights to its headnotes and its Key Number System.
Before training its AI tool, Ross sought licenses to Westlaw’s copyrighted content, but Thomson Reuters refused. Ross then entered an agreement with LegalEase to obtain training data in the form of “Bulk Memos.” Thes memos, created by lawyers instructed to use Westlaw headnotes, were sold to Ross. Thomson Reuters sued Ross for copyright infringement after discovering that Ross built its product using memos derived from Westlaw headnotes.
In 2023, Judge stefanos Bibas largely denied Thomson Reuter’s summary judgment motions. Though, he later stated that he “studied the case materials more closely and realized that [his] prior summary-judgment ruling had not gone far enough.” He continued the trial into 2024 and invited both parties to renew their summary judgment briefings. Both parties moved for summary judgment on fair use.
In February 2025, Judge Bibas revised his 2023 decision, granting summary judgment for Thomson Reuters against all of Ross’s copyright defenses, including fair use. He held that partial summary judgment should be granted on the direct copyright infringement claim for certain headnotes and denied Ross’s motions for summary judgment on direct copyright infringement and fair use.
Judicial Reasoning in 2025 Opinion
Judge Bibas persistent that Westlaw headnotes and the Key Number System met the Feist minimal threshold for originality and were thus copyrightable. He then analyzed the materials to determine if the two prongs for direct copyright infringement were met: actual copying and substantial similarity. Finding both, he concluded that direct copyright infringement occurred.
Ross invoked the affirmative fair use defense, which allows limited use of copyrighted materials without permission based on four statutory factors: 1) the purpose and character of the use, 2) the nature of the copyrighted work, 3) the amount of the work copied, and 4) the use’s effect on the existing and potential market. Judges have discretion in weighing these factors.
Fair Use Analysis in Thomson Reuters v. Ross
Factor 1: Purpose and character of the Use
The first factor frequently enough hinges on how “transformative” the use is. Judge Bibas ruled that Ross’s use was not transformative because it did not have a further purpose or different character from Thomson Reuters’s use. Because Ross used Westlaw headnotes as AI data to train a tool that answers legal questions, similar to Westlaw’s purpose, this factor favored Thomson Reuters.
Notably, Thomson Reuters’s headnotes did not appear in the Ross search user’s results but were copied at an intermediate step. Judge Bibas differentiated the Thomson Reuters case from computer programming cases where intermediate copyright has been permitted as fair use.
Factor 2: Nature of the Copyrighted Work
Judges analyze the degree of creativity of the original work, granting more protection to more creative works. Judge Bibas found that the headnotes contained the “minimal spark of originality” needed for copyright protection but were “not that creative.” This factor weighed in favor of Ross but was not as significant as factors #1 and #4.
Factor 3: Amount of the Copyrighted Work Used and Its Substantiality Relative to the whole Work
Judges consider how much of the whole work was copied.Judge Bibas held that this prong also weighed in favor of Ross because, although the intermediate training steps use a large amount and substantiality of the original work, the output does not. He stated, “Ross’s output to an end user does not include a West headnote.”
Factor 4: Impact on the Copyrighted Work’s Value or Potential Market
The final prong, how Ross’s use affected the copyrighted work’s value or potential market, also weighed in favor of Thomson Reuters. Judge Bibas considered this the most critically important factor, stating, “[t]he original market is obvious: legal-research platforms. And at least one potential derivative market is also obvious: data to train legal AIs.” He held that Ross did not show that such derivative markets do not exist and would not be impacted by the copying.
As Thomson Reuters prevailed on factors one and four, it ultimately prevailed on the overall balancing of the fair use factors.
Potential Impact on Other AI Copyright Cases
Thomson Reuters v. Ross is significant because defendants in other AI copyright cases have argued that their use of copyrighted materials is transformative fair use. In Thomson Reuters v. Ross, the defendant’s fair use argument failed, but future arguments could succeed as fair use is a fact-intensive analysis.
Judge Bibas stressed that this case only addresses a non-generative AI Tool. Because generative AI models could create outputs that are more transformative, the outcome of the first factor analysis could favor AI platforms in litigation involving generative systems.
Additionally, future jurisprudence in the AI-copyright field could follow a similar trajectory to software-related cases. In 2023, the Supreme Court stated in Warhol “a use that has a distinct purpose is justified as it furthers the goal of copyright, namely, to promote the progress of science and the arts, without diminishing the incentive to create.”
The other fair use factors may also come out differently in future fair use cases with different facts. In any event, Thomson Reuters vs. Ross demonstrates how fair use may be analyzed in the AI context going forward.
Beyond the fair use defense, Judge Bibas also rejected Ross’s other defenses, stating “[n]one of Ross’s possible defenses holds water. I reject them all.”
Business Takeaways
Copyright infringement liability in relation to AI is still uncertain. Parties should carefully consider how any given contract relating to AI allocates liability for potential copyright infringement. Customers of AI service vendors should review the scope of indemnities in their service agreements carefully. The Thomson Reuters v. Ross ruling might offer some leverage for customers when negotiating bespoke agreements and indemnity provisions.
For companies developing proprietary software that uses AI, it would be valuable to limit the use of third-party copyrighted content in training. Companies creating AI tools should consider each of the fair use factors in determining the type of content used to train their proprietary AI tools, and how the data is used.
it is important to remember that the fair use analysis is very fact-specific, and in the AI context, it will likely turn on the differences between purposes, type of content used, training methods, and outputs. Other AI developers might potentially be able to distinguish their cases from the facts of this case, and the outcome of generative AI copyright cases remains uncertain, but this case suggests that fair use likely will not shield all AI provider defendants.
AI Copyright Litigation: Thomson Reuters v. ross and the Fair Use Defense – Q&A Guide
Introduction
This thorough Q&A guide delves into the landmark AI copyright litigation case of Thomson Reuters v. Ross, examining its implications for AI providers and the future of copyright law. This case serves as a critical precedent, addressing the complex question of whether the fair use defense can protect AI model providers from copyright infringement claims when training AI models on copyrighted works.
General Questions
Q1: What is the significance of the Thomson Reuters v. Ross case?
The Thomson Reuters v. Ross case is significant as it is indeed the first major U.S. AI-copyright decision to address whether the fair use defense shields an AI model provider from copyright infringement claims. The court rejected Ross Intelligence’s fair use defense, asserting that using copyrighted data to train its AI model constituted direct copyright infringement and did not qualify as fair use based on the specific facts of the case. While outcomes may vary in other AI-copyright cases, this decision establishes a key precedent in copyright jurisprudence concerning AI and fair use.
Q2: What was the central issue in the Thomson Reuters v. Ross case?
The central issue was whether Ross Intelligence’s use of Thomson Reuters’ Westlaw headnotes and Key Number System to train its AI-powered legal research tool constituted copyright infringement or whether it was protected under the fair use doctrine.
Q3: What was the outcome of the Thomson Reuters v. Ross case?
The court granted summary judgment for thomson Reuters, finding that Ross Intelligence had infringed on Thomson Reuters’ copyrights and that the fair use defense did not apply.
Case Details
Q4: What are the facts of the Thomson Reuters v. Ross case?
In may 2020, Thomson Reuters sued Ross Intelligence, alleging copyright infringement. Ross developed a legal research search engine to compete with Westlaw which answered legal questions by providing relevant judicial opinions. Ross sought to use Westlaw’s headnotes and numbering system as a database to train its AI tool. Ross sought licenses to westlaw’s copyrighted content, but after Thomson Reuters refused, Ross entered an agreement with LegalEase to obtain training data in the form of “Bulk Memos” created by lawyers instructed to use Westlaw headnotes. Thomson Reuters sued Ross for copyright infringement after discovering that Ross built its product using memos derived from Westlaw headnotes.
Q5: What did Thomson Reuters allege in their complaint against Ross Intelligence?
Thomson Reuters alleged that Ross Intelligence intentionally infringed its copyrights by using Westlaw headnotes and Key number system to train its AI model.
Q6: How did Ross Intelligence obtain the data used to train its AI model?
After Thomson Reuters refused to grant Ross Intelligence a licence to use Westlaw content, ross obtained training data thru an agreement with LegalEase, which provided “Bulk Memos” created by lawyers instructed to use Westlaw headnotes.
Q7: What was Ross Intelligence’s argument in defense against the copyright infringement claim?
Ross Intelligence invoked the affirmative fair use defense, arguing that its use of the copyrighted material was transformative and thus permissible under copyright law.
Fair Use Analysis
Q8: What are the four factors considered in a fair use analysis?
the four statutory factors considered in a fair use analysis are:
1. The purpose and character of the use.
2. The nature of the copyrighted work.
3. The amount and substantiality of the portion used in relation to the copyrighted work as a whole.
4. The effect of the use upon the potential market for or value of the copyrighted work.
Q9: How did the court analyze the “purpose and character of the use” factor in Thomson Reuters v. Ross?
The court found that Ross’s use was not transformative as it did not have a further purpose or different character from Thomson Reuters’s use. Ross used Westlaw headnotes as AI data to train a tool that answers legal questions, similar to Westlaw’s purpose, which favored Thomson Reuters.
Q10: How did the court analyze the “nature of the copyrighted work” factor?
The court found that the headnotes contained the “minimal spark of originality” needed for copyright protection but were “not that creative.”,which leaned in favor of Ross,even though not as considerably as factors 1 and 4.
Q11: How did the court analyze the ”amount and substantiality of the portion used” factor?
The court held that this factor weighed in favor of Ross because, although the intermediate training steps use a large amount and substantiality of the original work, the output does not, stating, “Ross’s output to an end user does not include a West headnote.”
Q12: How did the court analyze the “impact on the copyrighted work’s value or potential market” factor?
The court resolute that it weighed in favor of Thomson Reuters, stating “[t]he original market is obvious: legal-research platforms. And at least one potential derivative market is also obvious: data to train legal AIs.” It held that Ross did not show that such derivative markets do not exist and would not be impacted by the copying and considered this the most critical factor.
Q13: Why did the fair use defense fail in the Thomson Reuters v. Ross case?
The fair use defense failed primarily because the court found that Ross’s use of the copyrighted material was not transformative and that it negatively impacted the potential market for Thomson Reuters’ copyrighted work.
Implications for AI and Copyright
Q14: What implications does the Thomson Reuters v. Ross case have for AI providers?
The decision casts doubt on whether the fair use defense will protect AI providers from liability when training their models on copyrighted works. AI providers should be aware that using copyrighted material to train AI models may constitute copyright infringement, especially if the use is not transformative and affects the market for the original work.
Q15: How might this case effect future AI copyright cases?
This case establishes a foundational precedent in copyright jurisprudence concerning AI and fair use. It may lead courts to scrutinize the transformative nature of AI’s use of copyrighted materials more closely and to consider the potential impact on derivative markets.
Q16: Will the outcome always be the same in other AI-copyright cases?
No, the outcome may differ in other AI-copyright cases, as fair use is a fact-intensive analysis.The specifics of the use, the nature of the copyrighted work, the amount copied, and the market impact will all be considered.
Q17: How does the court’s reasoning differentiate between generative and non-generative AI?
Judge Bibas stressed that this case only addresses a non-generative AI Tool. Because generative AI models could create outputs that are more transformative, the outcome of the first factor analysis could favor AI platforms in litigation involving generative systems.
Q18: What business takeaways can be derived from the Thomson Reuters v. Ross case?
Copyright infringement liability in relation to AI is uncertain, so parties should carefully consider how any given contract relating to AI allocates liability for potential copyright infringement.
Customers of AI service vendors should carefully review the scope of indemnities in their service agreements.
For companies developing proprietary software that uses AI, it would be valuable to limit the use of third-party copyrighted content in training.
Companies creating AI tools should consider each of the fair use factors in determining the type of content used to train their proprietary AI tools and how the data is used.
Summary Table: Fair Use Factors in Thomson reuters v.Ross
| Fair Use Factor | Court’s Analysis | Outcome |
| —————————————– | ——————————————————————————————————————————————————— | ——————– |
| Purpose and Character of the Use | Ross’s use was not transformative as it served a similar purpose to Westlaw. | Favored Thomson Reuters |
| Nature of the Copyrighted Work | Headnotes contained minimal originality, which favored Ross, but was not as significant as other factors. | Favored Ross |
| Amount and Substantiality of Portion Used | Intermediate training steps used a large amount and substantiality of the original work; however, the end product did not, which factore favored ross. | Favored Ross |
| Impact on Market | Ross’s use affected the potential market for Thomson Reuters’s services, especially in derivative markets like training legal ais. | Favored Thomson Reuters |
Conclusion
The thomson Reuters v. Ross case provides critical insights into the application of copyright law in the context of AI. While the fair use defense remains a complex and fact-specific inquiry, this case highlights the importance of considering the transformative nature of AI’s use of copyrighted materials and the potential impact on existing and derivative markets. AI providers must carefully evaluate their use of copyrighted content to mitigate the risk of copyright infringement claims.