Siyanna Lilova has posted Copyright or Copyleft for AI-Generated Works: Private Ordering Solutions for the Benefit of Content Creators on SSRN. Here is the abstract:
This dissertation attempts to find solutions that benefit authors for the up- and downstream AI challenges to international copyright law by taking inspiration from existing private ordering approaches. The work is structured in two main sections, one devoted to AI inputs and one to AI outputs, each exploring the main challenges for international IP law and possible solutions that can benefit content creators. First, the application of alternative compensation systems is explored as a possible solution to ensure both access to input data and provide fair remuneration for authors. With a view to enabling such a solution, the dissertation analyses how a global compensatory TDM [text and data mining] exception with mandatory remuneration for authors can be conceptualized within the international copyright framework and made compatible with the three-step test for copyright exceptions and limitations. Second, the paper analyses different regulatory approaches to the copyright protection of AI outputs, namely in the UK, US, EU, and Japan, and whether they answer identified copyright challenges related to generative AI. The paper concludes by arguing against the copyright protection of AI-generated works and explores how private ordering solutions can deal with the challenges for authors without the introduction of new exclusive rights.
