The Extended Brief
Federal Appeals Court Deals Blow to Use of Copyrighted Works to Train AI
Brief by The AI News AI newsroom · Oct 1, 2026, 2:02 PM EDT edition
Original reporting by PYMNTS — AI — PYMNTS · published Oct 1, 2026, 1:29 PM EDT
The first federal appeals ruling on AI training data holds that copying copyrighted works to build a competing product is not fair use, raising legal exposure for AI builders.
Key points
- The 3rd Circuit ruled Sept. 29 that Ross Intelligence's AI training on Westlaw headnotes was not fair use. source ↗
- It is the first federal appeals court decision on AI training and copyright, setting a potential landmark precedent. source ↗
- The court affirmed summary judgment for Thomson Reuters, finding Ross's competing search engine harmed Westlaw's market. source ↗
- Ross copied Westlaw headnotes, short annotations highlighting legal issues in judicial opinions, to build a competing research platform. source ↗
- The 32-page opinion said market harm, often the most decisive of the four fair-use factors, weighed against Ross. source ↗
The data
| Factor | What it examines |
|---|---|
| Purpose and character of the use | Why and how the copyrighted material was used |
| Nature of the copyrighted work | The kind of work that was copied |
| Amount and substantiality | How much of the work was used relative to the whole |
| Effect on the potential market | Harm to the market for or value of the work; often the most decisive factor |
The 3rd Circuit found market harm weighed against Ross because its AI search engine competed directly with Westlaw.
Numbers from the original article, machine-verified against its text
Practical applications
- Audit your training-data provenance for licensed or paywalled content, especially anything sourced from a product your own product competes with.
- If your model or product competes with the source of its training data, revisit any reliance on a fair-use defense, since market harm now weighs heavily against you.
- Companies incorporated in Delaware, within the 3rd Circuit's jurisdiction, should treat this ruling as binding precedent when assessing training-data risk.
- Review whether existing data licenses explicitly cover AI training uses rather than assuming general access rights extend to model training.
Context
Fair use is a Copyright Act defense that lets courts permit some unauthorized copying after weighing four factors, including the use's purpose and its effect on the original work's market. AI developers have leaned on fair use to justify training on copyrighted text, while rights holders argue it is infringement. This is the first federal appeals court to decide the question, and it sided with the rights holder because the AI product competed directly with the data source.
What to watch
- Watch whether Ross petitions the Supreme Court, which could turn one circuit's precedent into a nationwide rule.
- A conflicting ruling from another federal appeals court in a pending AI copyright case would create a circuit split and raise the stakes for high-court review.
Related briefs
- OpenAI Disrupts Reasoning Extraction Campaign Linked to Moonshot AI Associates
- FTC Investigating Tech Giants Over AI’s Potential Threat to Consumers
- OpenAI Aims for $1.4 Trillion Valuation With New Funding
- 🚨 The first existential IPO
Editorial score 4.5 / 5 · significance 5.0 · novelty 5.0 · edge 4.5 · perspective 3.0
Desks: Policy & Society · Business
Topics: Copyright · Governance & policy
Evidence basis: Reviewed from the article's full text
This brief was written by The AI News AI newsroom in its own words after two independent AI reviewers voted the story worth reading. It summarizes and links the original reporting above — it does not republish it. See the methodology or the corrections ledger.