TY  - GEN
N2  - Scholars have been grappling with the challenges posed by generative Artificial Intelligence (gAI) on copyright law. The question is whether we should apply 20th century laws to regulate 21st century technology, or should there be a recalibration of copyright law? In doing so, we cannot avoid the most fundamental question of all: what kind of culture do we want as a society? One diluted by gAI cultural products, where authors are replaced, or one where human authors get preferential treatment so that human creativity can continue to thrive?  Historically, human culture has been imperative to human flourishing, and has become arguably more sublime over time: from cave art to Da Vinci’s Mona Lisa, from oracle bone script to Shakespeare’s Macbeth. After social media seduced humanity into sharing what makes it tick (Douglas Rushkoff, ‘How We Taught Technology to Program Humans’, 18 February 2023, https://rushkoff.medium.com/how-we-taught-technology-to-program-humans-2388adf91364), gAI poses the threat of replacing human culture with increasingly diluted versions of that culture. If a foundational model is trained on products generated by AI, the quality of its output nosedives due to the ‘curse of recursion’ (Ilia Shumailov et al., ‘The Curse of Recursion: Training on Generated Data Makes Models Forget’ (2023), https://arxiv.org/abs/2305.17493). Then again, the ‘creativity of the machine’ could be increased by adding more randomness (‘increasing the temperature’ as they say in the jargon of computer science) and/or adding external sensory inputs. Human culture is a way in which humans idiosyncratically express themselves via works and share these with other humans. To outsource the evolution of human culture is to deprive humanity from actively shaping what it is like to be human and instead passively permeating society with a surrogate culture. To guarantee original creativity in the interim, society should provide preferential treatment to human creativity.
DO  - 10.1093/grurint/ikad133
DO  - doi
AB  - Scholars have been grappling with the challenges posed by generative Artificial Intelligence (gAI) on copyright law. The question is whether we should apply 20th century laws to regulate 21st century technology, or should there be a recalibration of copyright law? In doing so, we cannot avoid the most fundamental question of all: what kind of culture do we want as a society? One diluted by gAI cultural products, where authors are replaced, or one where human authors get preferential treatment so that human creativity can continue to thrive?  Historically, human culture has been imperative to human flourishing, and has become arguably more sublime over time: from cave art to Da Vinci’s Mona Lisa, from oracle bone script to Shakespeare’s Macbeth. After social media seduced humanity into sharing what makes it tick (Douglas Rushkoff, ‘How We Taught Technology to Program Humans’, 18 February 2023, https://rushkoff.medium.com/how-we-taught-technology-to-program-humans-2388adf91364), gAI poses the threat of replacing human culture with increasingly diluted versions of that culture. If a foundational model is trained on products generated by AI, the quality of its output nosedives due to the ‘curse of recursion’ (Ilia Shumailov et al., ‘The Curse of Recursion: Training on Generated Data Makes Models Forget’ (2023), https://arxiv.org/abs/2305.17493). Then again, the ‘creativity of the machine’ could be increased by adding more randomness (‘increasing the temperature’ as they say in the jargon of computer science) and/or adding external sensory inputs. Human culture is a way in which humans idiosyncratically express themselves via works and share these with other humans. To outsource the evolution of human culture is to deprive humanity from actively shaping what it is like to be human and instead passively permeating society with a surrogate culture. To guarantee original creativity in the interim, society should provide preferential treatment to human creativity.
T1  - Copyright as Affirmative Action for Human Authors Until the Singularity.
AU  - Friedmann, Danny,
JF  - GRUR International
VL  - 73, 1, 2024
LA  - eng
ID  - 49130
KW  - Copyright
KW  - Artificial intelligence.
KW  - Copyright.
KW  - Patents.
TI  - Copyright as Affirmative Action for Human Authors Until the Singularity.
LK  - https://doi.org/10.1093/grurint/ikad133
UR  - https://doi.org/10.1093/grurint/ikad133
ER  -