Patent and Trademark Office and the Copyright Office by surprise.
In its statement of interest—a formal declaration of the government's position in a private lawsuit—the DOJ argued that using copyrighted material to train artificial intelligence models should be considered "fair use." The department suggested that this practice is transformative and necessary to maintain the country's technological lead over foreign competitors.
This intervention is significant because while such filings are not legally binding, they often carry substantial weight in persuading judges on complex issues.
The DOJ's stance addresses a central conflict in AI infrastructure: whether AI firms must pay to license the massive datasets required to build large language models.
While the DOJ argued that enforcing copyright in this context would be overly cumbersome, it stopped short of deciding whether a formal licensing system for AI data is feasible.
The lack of coordination between the DOJ and federal copyright experts highlights a shift in how the government manages media and technology disputes.
According to sources, no career antitrust attorneys signed the filing, mirroring previous instances where senior officials acted independently of agency specialists.
This development directly impacts news publishers and AI developers, as it signals a federal preference for protecting the growth of the AI industry even when it conflicts with traditional intellectual property protections.