Image: Henrik Sorensen/DigitalVision/Getty Images.
Imagine you are a graphic designer and you decide to use Duet AI to generate images for a presentation to a client. A little later, you receive a message from an illustrator who claims that one of the images used in your presentation looks a lot like her work, and that she is suing you for copyright infringement.
Now, in a similar situation, Google has your back.
Google in the footsteps of Microsoft
Google has just announced that it will defend Google Cloud and Google Workspace users against intellectual property lawsuits related to the use of generative AI.
The place of copyright in the use of AI has been at the heart of debates for several months. That’s why Microsoft and other companies in the industry have made commitments similar to those of Google: last month, Microsoft announced its Copilot copyright commitment, promising to take responsibility for the legal risks associated with the claims. Copyright on AI-generated products using Microsoft Copilot services.
Liability subject to conditions
Thanks to this new service, users of Google Cloud and Google Workspace will benefit from compensation if a third party claims copyright in connection with training data or results generated by its AI. This is to protect against potential copyright infringement claims regarding the use of training data or content created by Google’s generative AI services.
This compensation will be granted on the condition that users respect certain guidelines relating to the responsible use of AI, for example by refraining from intentionally infringing the intellectual property of others. If a user attempts to create an image using AI by requesting a specific artwork or uploading copyrighted content, Google will not be held responsible for copyright infringement. author.
Google’s effort highlights the web giant’s desire to strengthen its partnership with its users, giving them some breathing space in their use of generative AI. As for giving relief to artists whose works have been used to train many of the large language models that underpin some generative AI tools, we’ll still have to wait.
Source: ZDNet.com