Major AI companies including OpenAI, Google, Meta, Microsoft, Alphabet, Stability AI, and IBM have been urged to compensate over 8,000 authors for using their copyrighted work to train generative AI models. In an open letter initiated by the US Authors Guild, the authors express concerns over plagiarism by AI systems that mimic and reproduce their language, styles, stories, and ideas without permission or compensation.
The letter emphasizes that generative AI technologies, particularly those built on large language models (LLMs), owe their existence to the writings of these authors. The authors argue that AI companies have benefited from millions of copyrighted books, articles, essays, and poetry without providing any form of compensation.
To address this, the Authors Guild hopes that settlements can be reached outside of court to avoid costly and time-consuming lawsuits. However, some authors have taken a more aggressive approach and filed lawsuits against those they believe have plagiarized their work.
This issue is not limited to the world of literature. In May, the Writers’ Guild of America (WGA) initiated a protest against the use of AI in scriptwriting for movies, referring to AI as plagiarism machines. The debate surrounding proper attribution and copyright in relation to generative AI technology continues to evolve and requires a resolution.
Companies such as OpenAI have constructed LLMs like GPT and the recent Llama-2 by scraping information from countless sources across the internet, including most websites. In an effort to address these concerns, OpenAI has entered into agreements with different organizations to access data for training their generative AI systems. Notably, they have partnered with the Associated Press, granting them access to text archives dating back to 1985, while OpenAI benefits from the agency’s technology and expertise.
It is essential to find a balance between the advancement of generative AI technologies and protecting the rights of authors and content creators. The ongoing struggle between attribution and copyright poses a challenge that must be addressed promptly.
By recognizing the value that authors bring to the development of AI systems, it is crucial for AI companies to seek consent and provide compensation for the use of copyrighted work. Resolving these issues outside of the courtroom can help foster a fair and equitable environment for authors, while enabling the continued progress of generative AI technologies.
In conclusion, the open letter serves as a call for collaboration between AI companies and authors to ensure that proper recognition and compensation are given for the contributions made by authors in the development of AI systems. By addressing these concerns, the industry can move forward and strike a balance that benefits both creators and innovators alike.
