Google recently confirmed that it will follow Microsoft's lead by limiting indemnity to customers of its generative AI services in the event of copyright infringement claims. This move comes as many generative AI services are being questioned for their reliance on material obtained from a variety of sources without requisite permissions or payments.
This is an issue of particular concern for businesses considering the use of generative AI for its increased risk of legal exposure. Microsoft was the first tech-giant to offer protection in an announcement stating it would defend its customers using products such as Copilot against any copyright infringement lawsuits.
Google has made a similar commitment. The company's VP Legal, Neal Suggs, and VP, TI Security & CISO at Google Cloud, Phil Venables, jointly stated, “If you are challenged on copyright grounds, we will assume responsibility for the potential legal risks involved.”
Cover for Duet AI and Vertex AI Users
Google's indemnity offer specifically caters to its users employing Duet AI, a chatbot integrated with Google Workspace apps, and those making use of the Vertex AI platform, which comprises AI Search, AI Conversation, AI Text, Multimodal Embeddings, and Visual Captioning software. In addition, the indemnity policy also covers users of code-generating Codey APIs.
“We are Google provides indemnity to Duet AI, Vertex AI, and Codey API users, safeguarding against claims of AI training and output infringing copyrights., as opposed to just the output argument,” a Google spokesperson confirmed to The Register. However, this protection will not extend to instances where a user intentionally prompts a model to generate content copying someone else's work directly.
Clearer Guidelines for Users
In light of the announcement, users are reminded of their obligations to use Google's AI services responsibly. The indemnity will only be applicable if the user has not intentionally created or used the generated output to impact other's legal rights. If users are found to be in breach, Google's legal team may not assist in their defence.
The company has announced it will review all relevant information with a customer for every copyright allegation and has encouraged users to use existing and emerging tools, such as citing resources, to use generated output responsibly. Despite these steps towards indemnification, the legal boundaries regarding AI systems violating copyright laws at the training or output stages remain to be clarified by the courts.