Source: Proskauer Rose LLP

Proskauer Rose LLP published an advisory on December 15, 2025, focusing on the critical importance of precise drafting in data license agreements in light of the generative AI boom. The article argues that standard limitations on "commercial use" or "derivative works" are no longer sufficient to prevent data from being ingested into Large Language Models (LLMs). It emphasizes that licensors must now include explicit restrictions regarding machine learning training, text and data mining (TDM), and the creation of synthetic data to protect their intellectual property assets.

The piece provides practical guidance for legal counsel, suggesting specific language to define "Artificial Intelligence purposes" broadly enough to cover future technological iterations. It also warns licensees to carefully review inbound data agreements to ensure they have the necessary rights to use datasets for model training, avoiding costly breach of contract lawsuits. The post highlights that as AI litigation evolves, the specific wording of license grants and restrictions will be the primary battlefield for determining the legality of model training activities.