Thomson Reuters has finally won its fight against the startup Ross Intelligence. In its complaint filed in 2020, the news agency accused it of using protected content from Westlaw — Thomson Reuters’ legal platform — to train its AI, without permission. The judge sided with the plaintiff: “None of Ross’s possible defenses hold water. I reject them all,” said Stephanos Bibas.
First major legal battle against an AI
Ross Intelligence had invoked “fair use,” a principle of copyright law whose contours vary from one country to another. In the United States, this allows, in certain circumstances, the use of a protected work without the authorization of the rights holder. A use assessed according to four criteria: the purpose and nature of the use, the nature of the protected work, the quantity and importance of the extracts used, and the impact on the market.
While Ross Intelligence hoped to rely on fair use, the judge found that the company had failed on two of the four criteria and that the most decisive was the impact on the market. The start-up offered a direct commercial alternative to Westlaw, which weighed heavily in the decision. Thomson Reuters is obviously pleased with this decision; as for Ross, the complaint led to the fall of the company in 2021.
Beyond the Westlaw case, this judgment is the first major copyright case involving AI in the United States. A major decision that will certainly interest authors, publishers and all content creators whose production is used to train the AI models of large companies in the sector. Several complaints are currently being investigated against OpenAI. The latter regularly invoke "fair use" to avoid asking for permission to use this content, nor of course paying anyone anything.
Source: Wired
0 Comments