The Generative AI: A Threat or Breakthrough for Copyright?
OpenAI, along with generative artificial intelligence in general, is facing criticism for using internet data to train its algorithms, accused of infringing copyright by “copying” existing works. However, a judge’s decision has dismissed this view, arguing that AI generates content through a process of synthesis, not literal copying. Let’s delve into the arguments and the impact of this ruling on the future of AI and copyright.
OpenAI vs. Raw Story Media and Alternet Media: the origin of the dispute
In February, Raw Story Media Inc. and Alternet Media Inc., news companies, sued OpenAI, claiming that ChatGPT was reproducing their articles “textually or almost textually.” Both companies argued that OpenAI had violated the Digital Millennium Copyright Act (DMCA) in the model training process. They also sought monetary damages and requested OpenAI to stop using their works without authorization. However, Judge Colleen McMahon ruled in favor of OpenAI, stating that removing copyright details to train an AI does not necessarily constitute an infringement.
Copying or synthesizing? The nature of generative AI
One key argument is that generative artificial intelligence does not reproduce content literally but synthesizes it. Like the human brain recalls and reconstructs information without replicating it exactly, AI creates based on patterns and structures, not through direct copying. Given the vast amount of data on which models are based, the probability of a literal reproduction of a specific text is very low.
Copyright and the AI dilemma
Critics of generative AI argue that training models with internet-available content amounts to “intellectual property theft.” However, from a legal perspective, copyright focuses on the copying and distribution of original content. In the case of generative AI, the creation process is completely different, and copyright laws do not address this new form of synthesis. This suggests that the concept of infringement may need revision in light of new technologies.
Is a change in legislation necessary?
Generative AI sparks a debate about the limits of copyright and the need to update legislation to balance innovation with intellectual protection. Although controversial, its synthesis capability enriches general knowledge by creating original content derived from thousands of sources.
In summary, the initial ruling supports the idea that AI does not copy but creates. The discussion on copyright and technology is far from over, but for now, the law seems to be on the side of innovation.
