With the rapid rise of artificial intelligence (AI) systems, especially Generative AI, a new wave of challenges regarding intellectual property is emerging. Generative AI, like Chat GPT or DALL-E, can produce essays, poems, artwork within seconds, often imitating renowned artists and their styles with remarkable precision. The Generative AI Copyright Disclosure Act is an attempt to keep in check the evolution of Generative AI and protect the work of artists.
How do generative AI systems do it?
AI generative models are trained with vast amounts of data, including texts or images. During the training process, AI platforms identify patterns to create rules and predictions, which then allow them to respond to the prompt given to them. However, this training process may likely result in intellectual property infringement. Generative AI models use data scraped from the web that may include copyrighted content.
The aim of the bill
This legislation, introduced by Adam Schiff, would require AI companies to disclose what copyrighted materials have been used to train their generative AI models. In other words, AI companies would have to submit a detailed notice to the Register of Copyrights. This notice should contain the copyrighted works used in their AI training datasets. The Act introduces civil penalties for non-compliance, with fines of $5,000 for failure to submit the notice. This mechanism aims to ensure that AI developers take their obligations seriously and prioritize transparency in their use of copyrighted materials. In other words, this legislation wants to tackle the concerns about the infringement of intellectual property rights in the field of artificial intelligence.
The legislation wants to make more visible the use of copyrighted materials in AI training. It allows copyright holders to seek compensation and would create a more ethical and transparent AI development ecosystem. This act has garnered support from various creative industry organizations such as SAG-AFTRA, Writers Guild of America or the Recording Industry Association of America. They see this bill as a vital step in protecting their members’ rights in the digital age.
Impact on AI Innovation
However, developers may face practical difficulties in identifying and reporting all copyrighted works in large datasets, especially those which are based on web scraping. The additional administrative overload and legal liabilities introduced by the bill could drastically reduce the pace of AI innovation. Small companies are likely to struggle to comply, while larger tech giants will be able to adapt more easily. Some AI companies,that want to avoid these regulations altogether, could decide to relocate outside the US. This exodus could put the U.S. at a competitive disadvantage in the global race for AI leadership. Moreover, when copyright owners start denying the use of their works for AI training, it would result in a reduction of the data available therefore further hindering AI innovation.
As the technology continues to advance at an unprecedented pace, the Generative AI Copyright Disclosure Act of 2024 represents a crucial first step in addressing the complex relation between AI development and intellectual property rights. Collaboration between policymakers, AI developers, and creative communities will be essential. It would allow us to refine and implement effective regulations that strike the right balance between fostering innovation and protecting the rights of creators in the digital age. This new bill could have ramifications worldwide and change the global legal framework regarding AI.
Sources :
https://katzarov.com/evolving-ai-transparency-the-generative-ai-copyright-disclosure-act/
https://networkreadinessindex.org/the-future-of-intellectual-property-in-the-era-of-ai/
https://hbr.org/2023/04/generative-ai-has-an-intellectual-property-problem