Our Take
The amendment signals the Times is tightening its case after initial filing; the specifics of what was added will determine whether this shifts legal exposure for both companies.
Why it matters
This is the largest media company lawsuit against AI builders over training data. How courts rule on the Times' claims will set precedent for whether publishers can sue over unlicensed model training, affecting settlement leverage across the industry.
Do this week
Legal and compliance teams: flag this case for monthly updates through discovery and any settlements, as rulings will likely reshape AI training data policies within 18 months.
The Times Strengthens Its Case
The New York Times filed an amended complaint in its copyright infringement lawsuit against OpenAI and Microsoft, expanding its original claims filed in December 2023. The updated filing adds new allegations and refines the legal arguments about how the two companies used Times content to train their large language models without permission or compensation.
The Times has not publicly disclosed the specific additions in the amended complaint, and the full filing documents are not yet widely available in reporting. Court filings in copyright cases often include supplementary evidence, witness declarations, or refined legal theories that strengthen a plaintiff's position ahead of summary judgment or trial.
This move is standard civil litigation procedure. Amendments allow plaintiffs to respond to defendants' initial motions, incorporate newly discovered evidence, or clarify claims based on discovery exchanges. The fact that the Times chose to amend rather than refile suggests the core case structure remains sound but needed reinforcement.
Why Legal Exposure Matters for Both Sides
The Times lawsuit is the highest-profile copyright case filed against a major AI builder. Other publishers and authors have filed similar suits, but the Times brings institutional credibility, legal resources, and a massive dataset of work allegedly used for training. A favorable ruling for the Times would expose OpenAI and Microsoft to statutory damages (up to $150,000 per work infringed under U.S. copyright law) and could establish that generative AI training on copyrighted content without license or fair-use defense is indefensible.
Conversely, if OpenAI and Microsoft prevail on a fair-use theory, they could establish a broad safe harbor for AI training on web-scraped copyrighted material. Settlement negotiations are already expected to pivot on the strength of the amended filing; publishers watching this case are likely using it to benchmark their own negotiating positions with AI labs.
The amendment timing matters. If the Times has uncovered new evidence of deliberate reliance on Times content (e.g., internal emails discussing the value of news articles for training), that hardens the case significantly. If the amendment is largely procedural, it signals the Times is preparing for a lengthy discovery process.
What This Means for Teams Building on LLMs
Companies using OpenAI and Microsoft models for production should not assume training data disputes are settled. If either company faces material licensing liability, it could affect their service costs, model availability, or terms of use. Teams should document what data their own applications train on and whether they have obtained or can defend usage rights.
For organizations licensing third-party content to build fine-tuned models, this case reinforces the importance of explicit agreements with rights holders. The Times litigation is pushing toward a regime in which AI builders will need either licenses, fair-use precedent, or explicit consent for copyrighted training material. Until a court rules definitively, assume that licensing is the safer path.