→ MIT 6.S979 Values and AI: Accidents, Alignment, and Misuse
December 2023
What is “fair”?
Reevaluating Copyright in
the
World of
Generative AI
Following the launch of OpenAI’s ChatGPT and StabilityAI’s Stable Diffusion models in 2022, there have been a number of lawsuits filed against generative AI companies that accuse them of copyright infringement. This paper establishes the scope of the fair use doctrine of the Copyright Act through a number of legal decisions and explores copyright implications at various stages of generative AI model development. This paper argues that past legal precedent about copyright law is not sufficient to address new issues presented by AI technology and presents possible policy approaches, particularly with regards to data licensing, that could address legal and ethical issues surrounding data usage during the development of these technologies.
hello
I. Introduction
gello
The launch of OpenAI’s ChatGPT and StabilityAI’s Stable Diffusion models in 2022 have made machine learning (ML) and artificial intelligence (AI) tools easily accessible to the general public. As a result, there has been an increase in the discussion of the legal implications of these models, especially relating to the use of copyrighted works to train these models and whether the outputs of these models could be considered copyright infringement.
There have been a number of lawsuits filed against generative AI companies that accuse them of copyright infringement. American stand-up comedian and actress Sarah Silverman filed a copyright lawsuit against OpenAI for unauthorized use of her memoir to train ChatGPT [14]. A group of programmers filed a class action lawsuit against OpenAI and Github for using their code as training data without attribution [15]. Getty Images filed a lawsuit against StabilityAI for training on 12 million copyright protected works [16].
These disputes fall within a larger set of related concerns about the development, applications, limitations, and consequences of AI technology, not only amongst academics, but also the government and the general public. The growing concerns about the ethical implications of these models and the companies who develop them have led to calls for regulation in the AI sector. In May 2023, the Writers Guild of America (WGA) went on strike to renegotiate contracts with major studios; amongst their requests were protections against studios’ use of AI so that writers’ compensation could not be unfairly reduced [21]. In June 2023, the EU AI Act was released, marking one of the first pieces of AI legislation in the world [17]. In July 2023, the Screen Actors Guild-American Federation of Television and Radio Artists (SAG-AFTRA) joined the WGA strike after contract negotiations broke down over issues relating to protections against AI [22]. The agreements reached at the end of the WGA and SAG-AFTRA strikes in the Fall of 2023 mark some of the first pieces of regulation imposed in any ...read more.