Aller au contenu

Cookies de mesure d’audience

Nous utilisons Google Analytics et Google Tag Manager pour compter les visites et voir quelles pages sont utiles. Si vous acceptez, Google Analytics peut déposer des cookies pour reconnaître vos visites suivantes. Si vous refusez, il ne dépose aucun cookie, mais Google reçoit tout de même un signal sans cookie pour chaque page que vous consultez. Nous n’envoyons pas à Google votre nom, votre adresse e-mail ni les informations de votre inscription, et vous pouvez modifier votre choix à tout moment via « Paramètres des cookies » en bas de la page. Confidentialité

Cette traduction de l'interface est un brouillon qui n'a pas été relu par une personne de langue maternelle. L'anglais est la version source. English

Third Circuit upholds Thomson Reuters' win over Ross Intelligence, rejecting AI training fair use defense

29 sept. 2026 · Collecté automatiquement à partir de sources publiques. Non vérifié par une personne.

Ce contenu n'est disponible que dans sa langue d'origine (anglais).

What happened: On 29 September 2026, the US Court of Appeals for the Third Circuit affirmed a Delaware district court's partial summary judgment for Thomson Reuters, holding that defunct legal-research company Ross Intelligence infringed copyright by using thousands of Westlaw's editorial headnotes to train its AI-powered legal search engine. The three-judge panel (opinion by Circuit Judge Tamika Montgomery-Reeves) rejected Ross's fair use defense under Section 107: the use was highly commercial and minimally transformative because both companies used the headnotes for the same purpose -- helping researchers find judicial opinions -- and Ross's tool retrieved existing legal texts rather than generating new content. The court's full reasoning is currently sealed. Ross shut down its platform in 2021, citing the cost of this litigation.

Why it matters to creators: Reuters describes this as the first ruling by a US appeals court in the wave of copyright cases over AI training. For creators, the key signal is that direct commercial competition weighs heavily against fair use, and the court adopted a narrow reading of 'transformative use' -- feeding protected material into an intermediate database for machine learning is not automatically transformative.

What is not yet known: The sealed reasoning limits what other courts can take from the decision, and the Third Circuit explicitly distinguished Ross's retrieval tool from foundation models that generate new content, so the ruling offers limited guidance for the generative-AI cases (authors, news, music) still pending.

Sources