Law:
federal law: big stuff
state law: small stuff
local law
Civil Law: no jail law
Criminal Law: jail law, with rights to jury trail
Queen Victoria's sketch case.
Andy Warhol's case of transformative use of portrait (Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith)
Frank Easterbrook: since there is no law for horse, we only need to apply old law to new technology. But one of the challenge about it is that algorithm might not behave as intended by the algorithm creator.
Laura first introduced three layers of law: federal law, state law, and local law. Intuitively, federal law takes care of big stuff directly under constitution like civil rights and social security. The state law takes care of smaller and region specific stuff like . Where as local law. In addition to three layers, Laura introduced the classification between Criminal and Civil Law. Intuitively speaking, Criminal Law . Civil Law .
Laura then introduced couple famouse cases that made the foundation of copyright law in U.S. Queen Victoria's sketch case _. Andy Warhol's case of transformative use of portrait (Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith) .
Laura first delineated three tiers of the legal framework: federal law, state law, and local law. She expounded that federal law primarily addresses overarching issues under the jurisdiction of the United States Constitution, such as civil rights and social security. In contrast, state law is tailored to address more localized and region-specific matters, such as disaster recovery of electricity grids. Local law, on the other hand, focuses on ordinances and regulations pertinent to smaller communities, often dealing with zoning, local safety, and municipal governance. Therefore, copyright law is more associated with state law. Furthermore, Laura elucidated the dichotomy between Criminal and Civil Law. Criminal Law is concerned with actions that are proscribed by statutes and constitute offenses against the state or the public at large, often resulting in penalties like imprisonment or fines. Civil Law, conversely, primarily deals with disputes between individuals or organizations, typically resolving in the form of monetary compensation or specific performance.
In her presentation, Laura delved into pivotal legal precedents that have significantly influenced the landscape of copyright law in the United States. She first discussed the historical context of Queen Victoria's sketches, which, although not a formal legal case, provide a crucial understanding of copyright expiration. Queen Victoria, during her reign in the 19th century, created numerous private sketches. Appropriation and reproducing these sketch sparked the creation of U.S. copyright laws.
Laura then shifted focus to a landmark legal battle, Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith. This significant case revolved around the legal concept of transformative use within the scope of fair use. The case involved a dispute between the Andy Warhol Foundation and photographer Lynn Goldsmith regarding a series of silk-screen portraits of the musician Prince, created by Andy Warhol, which were based on a photograph taken by Goldsmith. The crux of the case lay in determining whether Warhol’s adaptations were sufficiently transformative of Goldsmith’s original photograph to qualify as fair use, thereby not infringing on her copyright. This case serves as a pivotal example in understanding how derivative works are assessed under copyright law, particularly in artistic adaptations, setting a precedent for future interpretations of transformative use in copyright disputes.
In the class activity, Laura showed us cases of prompting diffusion models might unintentionally generate copyrighted images, which proves the misuse of copyrighted image in training dataset of diffusion models. Then we were asked to produce copyrighted image without explicitely mentioning the subject in the images. In the second class activity, we were asked to create charts of ban / high risk / low risk use of AI technology. Interestingly, while all students agree military applications should ban the use of AI, the government agent's answer completely exclude this use. This might due to military have a separate law. New York, California, Illinois, Maryland all took some move on ensuring workers are hired fairly without discrimination based on gender, race, and ethnicity.
In the follow up discussions, we talked about how copyright grants from AI company is actually enforceable by law. Our discussion come to a conclusion that whether a generated image's copyright can be granted is solely based on the image itself regardless of underlying training data.
Laura explained the legal framework in the U.S. as consisting of three levels: federal, state, and local law. Federal law, governed by the U.S. Constitution, deals with national issues like civil rights, while state law focuses on regional concerns, such as disaster management for electricity grids. Local law is concerned with community-specific matters like zoning and local safety. She noted that copyright law is typically associated with state law. Additionally, Laura differentiated between Criminal Law, which deals with offenses against the state and often results in penalties, and Civil Law, which resolves disputes between parties, usually through monetary compensation or specific performance.
In her presentation, Laura discussed significant legal precedents in U.S. copyright law. She began with Queen Victoria's 19th-century sketches, which, although not a formal legal case, helped shape the understanding of copyright expiration. She then examined the case of Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith, a landmark lawsuit that focused on the concept of transformative use in fair use. This case, involving a dispute over Warhol's Prince portraits based on Lynn Goldsmith's photograph, set a precedent for how derivative works, especially in art, are evaluated under copyright law.
In a class activity, Laura demonstrated how diffusion models might inadvertently generate copyrighted images, suggesting misuse of copyrighted material in their training datasets. We were then asked to create copyrighted images without explicitly mentioning the subject. In another activity, we created charts categorizing AI technology uses as banned, high risk, or low risk. Interestingly, all students agreed on banning AI in military applications, in contrast to a government agent's view, likely due to separate military laws. We also discussed moves by states like New York, California, Illinois, and Maryland to ensure fair hiring practices without discrimination.
Finally, we discussed the enforceability of copyright grants from AI companies, concluding that the copyright eligibility of a generated image depends solely on the image itself, not the training data used to create it.
Table of Content