protecting artists against AI: copyright

Examining AI under copyright law is currently difficult, as it is largely uncharted territory. Most notably, the question of whether or not AI is considered fair use is in play. Within the US, the act of data mining and scraping from the internet is not considered illegal; it is considered fair use. Moreover, each case of generative AI usage has different circumstances and may be judged differently on being a transformative work. Whether or not a work is transformative depends on a number of factors, including if it is commercial or non-commercial. Because generative AI aims to compete with the same market its data is from, the fourth factor of fair use (the effect of the use on the market) is called into play. AI usage must be looked at on a case-by-case basis in order to find whether or not they are negatively affecting the market using the artists’ own data.
In the EU, the GDPR (General Data Protection Regulation) system allows artists to license their work and give or deny consent for its use. However, the management of such a licensing and consent system would be inefficient and impractical, to say the very least. With the sheer scale of AI models, it is virtually impossible to contact and manage consent from every contributor. This leaves companies without a coherent and ethical system of obtaining all of their needed data.
Overall, many artists are morally opposed to their works being scraped by AI models. Companies are profit-focused, leading them to take offers from AI companies if the conditions create more profit. While artists and large companies are in opposition, their motivations differ, again lending to the idea that copyright is not an efficient route to take. Thus, in the law, different ways must be found to protect artists.

Image by Allison Vu from allisonvu.com