Summary
Nvidiahas found itself on the receiving end of a class action lawsuit alleging copyright infringement by its artificial intelligence platform NeMo. The litigation poses an obstacle toNvidiafully realizing its AI ambitions that have propelled its stock to unprecedented heights.
The Santa Clara, California-based tech giant has been trying to stay at the forefront of the AI revolution in more ways than one. A significant portion of its recent research and development efforts were hence focused on producing and optimizing hardware for machine learning applications. At the same time, the company has also been trying to innovate on the software side of things, oftentimes with great success, as underlined by the high adoption rate of technologies likeNvidia’s DLSS solution for upscalingand (as of version 3.5) frame generation.
Nvidia’s Generative AI Targeted by Copyright Infringement Lawsuit
Its ongoing artificial intelligence push has also expanded to generative AIs as of a few years ago. The company’s flagship offering in this niche is NeMo, an end-to-end cloud framework for building and deploying generative AI models akin to ChatGPT. This solution has now been targeted by a class action lawsuit initiated by a number of novelists. As first reported byReuters, authors Abdi Nazemian, Brian Keene, and Stewart O’Nan claim that their works were part of NeMo’s training dataset meant to help teach it how to simulate human writing. Since their works are said to have been included in the training data without their permission, the novelistsallege copyright infringementon Nvidia’s part.
The dataset at the center of the lawsuit reportedly comprised over 196,000 books as of October 2023, when it was taken down following a complaint by the copyright holders. The plaintiffs are seeking damages for the alleged unlicensed use of their works for commercial purposes, though the extent of their compensation target is currently unclear.
On a fundamental level, the newly filed litigation is similar to the classlawsuit that George RR Martin and many other authors filed against OpenAIin September 2023. Both cases involve creatives alleging that the use of their work to train generative AI models constitutes commercial use, which requires explicit permission under the DMCA. Meta and Microsoft, which is one of OpenAI’s investors, have been targeted by similar litigation in recent years.
Stateside courts have yet to set a specific precedent when it comes to generative AI specializing in writing. But creatives did already score a win in August 2023, when a U.S. court ruled thatAI-generated artcannot be copyrighted precisely because the models used to create it are trained on already copyrighted materials. While the same technically holds true for human artists and all art is ultimately derivative, the root of the issue is that generative AI models can recall and reproduce their training data with perfect accuracy, thus muddling the distinction between remixed and plagiarized works.
