A Sui Generis system solves generative AI IP better than copyright
Bonadio and McDonagh 20 [Enrico and Luke; June 26; Enrico Bonadio is a reader at City, University in London. He teaches, researches, and advises in the field of IP law; Luke McDonagh is an Associate professor at LSE law school; SSRN; “Artificial Intelligence as Producer and Consumer of Copyright Works: Evaluating the Consequences of Algorithmic Creativity”;p.12 //ST]
We have seen that at present copyright regimes may not be fit to accommodate what is produced by algorithmic creativity; indeed, copyright laws in several jurisdictions do not explicitly protect machine created works. The lack of human authorship/originality is a key factor in this regard. Yet, whether we agree or not, it is inevitable that corporate entities will lobby for a form of exclusive rights to protect the final outputs of machine-driven processes. Could an acceptable compromise be offered by a sui generis system—a kind of protection that could incentivise the development of and use of AI creative platforms while at the same time safeguarding human ingenuity?185 The benefit of a sui generis regime (as opposed to using the full scope of copyright to protect such works) would be that right holders could be given only a thin scope of protection, allowing them to prevent others from exploiting exact copies of the machine-generated work. In this view, it would essentially be protection against literal copying only.186 As to the length of protection, unlike the typical copyright duration, a very short duration could be applied in the AI context e.g. three years from the date of publication of the work (the length of term suggested in a recent AIPPI “Study Question” by Dutch delegates).187 Why should a thinner and shorter right be preferred?188 The usual justifications for the long duration in the context of human authors would not apply. In fact, providing AI developers and companies incentives for AI-created works by offering the strong traditional copyright protection/duration may even lead to fewer human-generated works being created in the long run.189 AI creative capacity is potentially both more vast and speedier than human capacity.190 The risk is a devaluation of human intellectual ingenuity and a marginalisation of the human creative potential.191 Just as automation threatens (and, over time, eliminates) existing jobs in manufacturing, AI creativity could threaten the value of human authorship. The provision of a limited sui generis right might neutralise this risk by providing an incentive to recognise the creativity of AI works without giving machines an equal level of protection to humans.192 The fact that only literal copying would be prohibited would leave human creators free to adapt, transform and reinterpret AI-generated works and thus to use them for creative purposes. In fact, such a feature of the proposed sui generis right may fit with the characteristics of many AI machines such as Amper, an AI music composition system.