The News Media Alliance (NMA) asserts that in order to train their models, artificial intelligence companies largely rely on the illicit extraction of copyrighted content from journalists and news outlets. The NMA claimed that the data sets used to train AI models contain much more news publisher material than other sources in a 77-page white paper and accompanying filing to the US Copyright Office. As a result, news outlets face competition from AI models and copyright violations by AI-generated content.
The NMA contends that although news organisations make investments and take chances, AI developers gain from users, data, the development of brands, and advertising revenue. Publishers also confront diminished earnings, job openings, and damaged connections with viewers. The NMA suggested that the Copyright Office declare that it is detrimental to publishers to use a publication’s content to monetize AI systems in order to address these concerns. In order to prevent the absorption of copyrighted materials and to take action to remove protected content from websites owned by third parties, the group also demanded a variety of licencing models and transparency requirements.
The NMA stated that although generative AI has advantages, publishers and journalists can employ AI for idea development, proofreading, and search engine optimisation. Over the past year, there has been a rise in the usage of AI chatbots such as Claude from Anthropic, Google’s Bard, and OpenAI’s ChatGPT. All of the techniques used to train these AI models have drawn criticism, and copyright infringement lawsuits have been filed against them. Though its Bard search tool is not covered by this legal protection commitment, Google has declared that it will take legal responsibility if its users are accused of violating copyright by utilising its generative AI products on Google Cloud and Workspace. A request for comment was not immediately answered by Google or OpenAI.