Google Pledges to Assume Legal Risks for Users of Its Generative AI Products Regarding Copyright Issues
-
Google has pledged to assume legal risks for users of its generative AI products to address copyright infringement lawsuits. The company explicitly stated that users of products now integrated with generative AI features will receive legal protection, aiming to alleviate increasing concerns about potential copyright violations. Google specifically mentioned seven products covered by this legal protection, including Duet AI in Workspace, Vertex AI Search, Vertex AI Conversations, Vertex AI Text Embedding API, and Vertex AI Visual Captions. However, Google's Bard search tool was not included.
Image credit: AI-generated image, licensed by MidjourneyGoogle stated, "If you are challenged on copyright grounds, we will assume the potential legal risks." The company described this as a "two-pronged, industry-first approach" to intellectual property indemnification, covering both training data and outputs from foundational models. This means if someone is sued because Google's training data includes copyrighted material, Google will bear the legal responsibility.
Google noted that indemnification for training data is "not actually a new protection." However, the company acknowledged that users want clear confirmation on whether their safeguards extend to cases where training data might contain copyrighted information.
Protection will also apply if users face lawsuits due to outputs generated by foundational models—for example, if they produce sentences resembling published works. Google clarified that this protection "only applies if you did not intentionally create or use generated output to infringe others' rights."
Other companies have made similar commitments. Microsoft announced it would assume legal responsibility for enterprise users of its Copilot products. Adobe stated it would shield enterprise customers using Firefly from copyright, privacy, and publicity rights claims.
Copyright issues have been troubling generative AI platforms, with more infringement lawsuits being filed against various companies. The most recent lawsuit was brought by renowned authors such as George R.R. Martin, John Grisham, and Jodi Picoult, among others. Additionally, Google faced a proposed class-action lawsuit alleging the use of personal information and copyrighted data to train AI models.