Skip to main content

Google promises to take the legal heat in users’ AI copyright lawsuits

Google promises to take the legal heat in users’ AI copyright lawsuits


The company says it will take responsibility for its training data and the output of its foundation models.

Share this story

If you buy something from a Verge link, Vox Media may earn a commission. See our ethics statement.

Photo illustration of a gavel casting a shadow over the Google logo
Illustration by Cath Virginia / The Verge

Google will protect customers who use some of its generative AI products if they get sued for copyright infringement, the company says.

In a blog post, Google said customers using products that are now embedded with generative AI features will be protected, attempting to assuage growing fears that generative AI could run afoul of copyright rules. It specifically mentioned seven products it would legally cover: Duet AI in Workspace (including text generated in Google Docs and Gmail and images in Google Slides and Google Meet), Duet AI in Google Cloud, Vertex AI Search, Vertex AI Conversation, Vertex AI Text Embedding API, Visual Captioning on Vertex AI, and Codey APIs. Google’s Bard search tool was not mentioned.

“If you are challenged on copyright grounds, we will assume responsibility for the potential legal risks involved,” the company said.

Google said it will follow a “two-pronged, industry-first approach” for intellectual property indemnification, which will cover its training data and results created from its foundation models. This means if someone gets sued because Google’s training data used copyrighted material, Google will take that legal heat.

The company said indemnity around training data “is actually not a new protection.” But Google admitted customers wanted explicit clarification that its protection covers the possibility that the training data took in copyrighted information. 

Google will also protect users if they are sued for the results they get after using its foundation models. For example, if they generate a sentence similar to a published work. The company noted this protection “only applies if you didn’t try to intentionally create or use generated output to infringe the rights of others.”

Other companies have made similar proclamations. Microsoft announced it will take legal responsibility for enterprise users of its Copilot products. Adobe said it would protect enterprise customers using Firefly over copyright, privacy, and publicity rights claims.  

Copyright issues have haunted generative AI platforms, and more lawsuits have now been filed against different companies for allegedly infringing on copyright. One of the latest lawsuits was filed by famous authors like George R.R. Martin, John Grisham, and Jodi Picoult.

Google already faced a proposed class action lawsuit for allegedly taking personal information and copyrighted data to train AI models, Reuters reported