TL;DR:
- The US Copyright Office has initiated a public comment period from August 30 to November 15, delving into the intersection of generative AI systems and copyright laws.
- Four key areas of focus include the use of copyrighted materials to train AI models, the copyrightability of AI-generated content, liability in cases of AI-generated content infringement, and the impact of AI emulating human voices and styles.
- Generative AI tools have rapidly evolved, prompting copyright disputes and legal actions, prompting the Copyright Office to address these concerns.
- Notable instances include copyright registration for AI-generated content, legal disputes against AI companies, and efforts to block data scraping for AI model training.
- Copyright regulations for AI-generated art are nuanced, and the threshold of human involvement for copyright eligibility remains a question.
- The Copyright Office seeks public input on the use of copyrighted materials in training datasets for AI models.
Main AI News:
In a move reflecting the fast-paced evolution of technology, the US Copyright Office has embarked on a journey to delve into the intricate interplay between generative AI systems and copyright statutes. Commencing on August 30, a public comment phase has been set in motion to probe the labyrinthine terrain where artificial intelligence intersects with intellectual property rights. This enlightening period of discourse will conclude its deliberations on November 15. The insights gleaned from these deliberations could potentially steer the direction of future copyright dispositions.
Documented across a 24-page dossier, available as a downloadable PDF on the Federal Register’s platform, the “Notice of inquiry and request for comments” poses inquiries that resonate with profound implications for the landscape of creative ownership in the United States.
Four distinct domains have captured the attention of the Copyright Office: Firstly, the inquiry navigates the realm of utilizing copyrighted content to train AI models, raising questions on whether such deployment constitutes an infringement. Secondly, a focal point emerges around the extent to which AI-generated outputs should be eligible for copyright protection, particularly when human intervention guides the operations of the AI system. Thirdly, a contemplative exploration into the contours of liability ensues, particularly when AI-generated creations encroach upon existing copyright privileges. Lastly, the profound influence of AI in mirroring human artistic voices and styles is acknowledged; although not strictly a copyright concern, it does traverse the sphere of state laws related to publicity rights and fair competition.
This solicitation for comments has been prompted by the advent of generative AI tools, which, in a remarkably brief span, have showcased remarkable capabilities, spanning from image and video rendering to voice emulation, query response, and on-demand text composition. Amidst the dazzle and skepticism surrounding these AI models, replete with corporate investments and legal actions, the Copyright Office finds itself at a critical juncture. The rise in AI-centered copyright disputes has thrust the agency into heightened vigilance:
Over recent years, the Copyright Office has fielded applications for registering works that encompass AI-generated content, with some applications attributing authorship or co-authorship to AI systems. Simultaneously, copyright holders have initiated claims of infringement against AI enterprises, rooted in the methodologies employed in training and the subsequent outputs generated by generative AI systems.
With apprehensions about the ascent and ambiguities accumulating, both the Congress and the Copyright Office have been engaged by a spectrum of stakeholders, each representing diverse viewpoints. An expansive endeavor was announced by the Office earlier this year, signaling their commitment to comprehensively explore these matters. The current “Notice of inquiry” stands as a manifestation of this initiative, drawing from the Office’s reservoir of research, expertise, and prior undertakings, augmented by invaluable insights contributed by stakeholders.
Highlighted within the contours of the Copyright Office’s dossier are instances previously chronicled, including the registration of latent diffusion artwork under Kris Kashtanova’s name, which was granted in September only to be conditionally revoked in February. Another noteworthy instance pertains to an endeavor to bestow sole authorship to an AI-created artwork by Stephen Thaler, akin to his pursuits in the realm of patents, which was met with judicial refusal. Further, a legal skirmish involving OpenAI is noted, where the company faced a copyright lawsuit from Sarah Silverman, pertaining to her copyrighted works’ incorporation within ChatGPT’s training data. Notably, artists have joined forces in a class action lawsuit in January against Stability AI, Midjourney, and DeviantArt, citing alleged copyright transgressions.
In response to analogous copyright apprehensions, a cohort of news organizations, including Conde Nast, the parent company of Ars Technica, have recently taken steps to curtail OpenAI’s web crawler, with the hope of mitigating data scraping for training upcoming AI models. A prevalent practice among major language models, such as those empowering ChatGPT, involves absorbing information from openly accessible online documents, typically without consulting copyright proprietors.
Navigating the Nuances: Copyright Law in the Era of Emergent Tech
In the realm of copyright law, the framework remains unchanging, yet its interpretation in the context of nascent technologies is nuanced and dynamic, contingent upon regulatory and judicial perceptions, often influenced by public sentiment.
As it stands, copyright protection in the United States can extend to AI-generated art within the context of a larger, human-authored work. However, AI-generated creations as standalone entities are not afforded individual copyright privileges. Yet, unresolved queries loom over the threshold of human involvement requisite to trigger eligibility for copyright registration—a reminiscent scenario akin to the advent of the camera in the 19th century.
In the 1884 case of Burrow-Giles Lithographic Co. v. Sarony, the defendant contested the copyright eligibility of photographs, asserting that photos merely reproduced the features of natural subjects and lacked creative essence. However, the court ruled in favor of copyright protection for photographs, highlighting their role as representations of the original intellectual concepts of an author.
Today, the query assumes a modern guise: Which actions undertaken through machine learning tools embody the “original intellectual conception of an author?” Is it the selection of a specific AI tool? The input provided to the AI? The curation or editing of the AI’s resultant output? Despite existing US copyright guidelines, these queries remain uncharted territory, rendering the Copyright Office receptive to the public’s insights.
Moreover, the Copyright Office seeks perspectives on the contentious issue of integrating copyrighted content within training datasets. The query acknowledges divergent viewpoints on whether the use of copyrighted materials to develop datasets for AI training—both generative and non-generative systems—constitutes infringement.
But the ambit doesn’t conclude there. The initial 24-page document, laden with intellectual intrigue, encapsulates 34 multipart questions, each offering a glimpse into the Copyright Office’s agenda, poised for further elucidation in a document of equivalent length.
The calendar earmarks October 18 as the deadline for submitting written comments, channeled through the Regulations.gov platform. Subsequent responses to initial comments are expected to be submitted to the Copyright Office no later than November 15.
Conclusion:
The Copyright Office’s exploration of AI and copyright signifies a critical response to the rising influence of generative AI tools. The intersection of these technologies with copyright laws has significant implications for content creators, AI companies, and industries relying on intellectual property. Clear guidelines and regulations will be crucial for fostering innovation while protecting creative ownership rights.
Source
The US Copyright Office has initiated a public comment period, delving into the intersection of generative AI systems and copyright laws
TL;DR:
Main AI News:
In a move reflecting the fast-paced evolution of technology, the US Copyright Office has embarked on a journey to delve into the intricate interplay between generative AI systems and copyright statutes. Commencing on August 30, a public comment phase has been set in motion to probe the labyrinthine terrain where artificial intelligence intersects with intellectual property rights. This enlightening period of discourse will conclude its deliberations on November 15. The insights gleaned from these deliberations could potentially steer the direction of future copyright dispositions.
Documented across a 24-page dossier, available as a downloadable PDF on the Federal Register’s platform, the “Notice of inquiry and request for comments” poses inquiries that resonate with profound implications for the landscape of creative ownership in the United States.
Four distinct domains have captured the attention of the Copyright Office: Firstly, the inquiry navigates the realm of utilizing copyrighted content to train AI models, raising questions on whether such deployment constitutes an infringement. Secondly, a focal point emerges around the extent to which AI-generated outputs should be eligible for copyright protection, particularly when human intervention guides the operations of the AI system. Thirdly, a contemplative exploration into the contours of liability ensues, particularly when AI-generated creations encroach upon existing copyright privileges. Lastly, the profound influence of AI in mirroring human artistic voices and styles is acknowledged; although not strictly a copyright concern, it does traverse the sphere of state laws related to publicity rights and fair competition.
This solicitation for comments has been prompted by the advent of generative AI tools, which, in a remarkably brief span, have showcased remarkable capabilities, spanning from image and video rendering to voice emulation, query response, and on-demand text composition. Amidst the dazzle and skepticism surrounding these AI models, replete with corporate investments and legal actions, the Copyright Office finds itself at a critical juncture. The rise in AI-centered copyright disputes has thrust the agency into heightened vigilance:
Over recent years, the Copyright Office has fielded applications for registering works that encompass AI-generated content, with some applications attributing authorship or co-authorship to AI systems. Simultaneously, copyright holders have initiated claims of infringement against AI enterprises, rooted in the methodologies employed in training and the subsequent outputs generated by generative AI systems.
With apprehensions about the ascent and ambiguities accumulating, both the Congress and the Copyright Office have been engaged by a spectrum of stakeholders, each representing diverse viewpoints. An expansive endeavor was announced by the Office earlier this year, signaling their commitment to comprehensively explore these matters. The current “Notice of inquiry” stands as a manifestation of this initiative, drawing from the Office’s reservoir of research, expertise, and prior undertakings, augmented by invaluable insights contributed by stakeholders.
Highlighted within the contours of the Copyright Office’s dossier are instances previously chronicled, including the registration of latent diffusion artwork under Kris Kashtanova’s name, which was granted in September only to be conditionally revoked in February. Another noteworthy instance pertains to an endeavor to bestow sole authorship to an AI-created artwork by Stephen Thaler, akin to his pursuits in the realm of patents, which was met with judicial refusal. Further, a legal skirmish involving OpenAI is noted, where the company faced a copyright lawsuit from Sarah Silverman, pertaining to her copyrighted works’ incorporation within ChatGPT’s training data. Notably, artists have joined forces in a class action lawsuit in January against Stability AI, Midjourney, and DeviantArt, citing alleged copyright transgressions.
In response to analogous copyright apprehensions, a cohort of news organizations, including Conde Nast, the parent company of Ars Technica, have recently taken steps to curtail OpenAI’s web crawler, with the hope of mitigating data scraping for training upcoming AI models. A prevalent practice among major language models, such as those empowering ChatGPT, involves absorbing information from openly accessible online documents, typically without consulting copyright proprietors.
Navigating the Nuances: Copyright Law in the Era of Emergent Tech
In the realm of copyright law, the framework remains unchanging, yet its interpretation in the context of nascent technologies is nuanced and dynamic, contingent upon regulatory and judicial perceptions, often influenced by public sentiment.
As it stands, copyright protection in the United States can extend to AI-generated art within the context of a larger, human-authored work. However, AI-generated creations as standalone entities are not afforded individual copyright privileges. Yet, unresolved queries loom over the threshold of human involvement requisite to trigger eligibility for copyright registration—a reminiscent scenario akin to the advent of the camera in the 19th century.
In the 1884 case of Burrow-Giles Lithographic Co. v. Sarony, the defendant contested the copyright eligibility of photographs, asserting that photos merely reproduced the features of natural subjects and lacked creative essence. However, the court ruled in favor of copyright protection for photographs, highlighting their role as representations of the original intellectual concepts of an author.
Today, the query assumes a modern guise: Which actions undertaken through machine learning tools embody the “original intellectual conception of an author?” Is it the selection of a specific AI tool? The input provided to the AI? The curation or editing of the AI’s resultant output? Despite existing US copyright guidelines, these queries remain uncharted territory, rendering the Copyright Office receptive to the public’s insights.
Moreover, the Copyright Office seeks perspectives on the contentious issue of integrating copyrighted content within training datasets. The query acknowledges divergent viewpoints on whether the use of copyrighted materials to develop datasets for AI training—both generative and non-generative systems—constitutes infringement.
But the ambit doesn’t conclude there. The initial 24-page document, laden with intellectual intrigue, encapsulates 34 multipart questions, each offering a glimpse into the Copyright Office’s agenda, poised for further elucidation in a document of equivalent length.
The calendar earmarks October 18 as the deadline for submitting written comments, channeled through the Regulations.gov platform. Subsequent responses to initial comments are expected to be submitted to the Copyright Office no later than November 15.
Conclusion:
The Copyright Office’s exploration of AI and copyright signifies a critical response to the rising influence of generative AI tools. The intersection of these technologies with copyright laws has significant implications for content creators, AI companies, and industries relying on intellectual property. Clear guidelines and regulations will be crucial for fostering innovation while protecting creative ownership rights.
Source