Judge sides with Thomson Reuters in fair use copyright case against AI startup

• A Delaware judge has ruled that Ross Intelligence is liable for copyright infringement, marking a significant moment in the ongoing debate about AI training methods. The lawsuit, filed by Thomson Reuters, claims Ross used its Westlaw database without permission.
• Ross’s defense argued fair use allowed the company to use copyrighted material to train its AI.
• This ruling could have a significant impact on future cases involving companies like OpenAI and Microsoft.

---

Full Story

A Delaware judge has delivered a landmark ruling that could significantly impact the way artificial intelligence companies train their models. On Tuesday, Feb. 11, Thomson Reuters secured a major victory against an AI startup, Ross Intelligence, which was accused of using copyrighted work without permission.

Ross Intelligence liable for copyright infringement

Judge Stephanos Bibas ruled that Ross Intelligence is liable for copyright infringement, rejecting the company’s defense claims.

Bibas stated, “None of Ross’s possible defenses holds water. I reject them all.”

The lawsuit, filed in 2020, predates the ongoing debate over the use of copyrighted data by AI companies.

“We are pleased that the court granted summary judgment in our favor and concluded that Westlaw’s editorial content created and maintained by our attorney editors, is protected by copyright and cannot be used without our consent. The copying of our content was not ‘fair use.’“

Statement by Thomson Reuters spokesperson

Copyright law and fair use debate

Under U.S. law, copyright owners have control over their work, allowing them to sell and distribute it. However, the fair use doctrine provides an exception, allowing certain uses without permission, such as for research or teaching.

Thomson Reuters’ Westlaw database is a widely used tool that summarizes legal rulings. Ross Intelligence attempted to license Westlaw’s content but was denied due to being a direct competitor.

Court documents later revealed that Ross obtained its data from LegalEASE, a company using bulk memos containing information from Westlaw.

Ross’s cofounder publicly denied the allegations in a blog post, claiming the lawsuit was an attempt to hinder his company’s growth. Despite the defense, Ross announced months later it was shutting down its operations, saying the lawsuit had crippled the business.

Ross Intelligence argued its use of Thomson Reuters’ Westlaw database to train its AI models fell under fair use. However, the court disagreed, ruling the company’s actions did not meet the criteria for fair use.

Impact on future cases

This decision marks the first ruling of its kind regarding fair use and AI, which could influence future legal battles surrounding AI training methods. The case is particularly notable given its timing, potentially impacting the rise of AI systems like ChatGPT.

The ruling could have significant ramifications for high-profile cases, such as the ongoing lawsuit between OpenAI and The New York Times, which accuses OpenAI of using its copyrighted content to train its models.

While the judge’s decision on copyright infringement is a major development, the case is not over. The matter will still proceed to a jury trial to determine the full extent of the damages. Ross Intelligence did not respond to requests for comment regarding the ruling.