Date of Award

9-2023

Document Type

Dissertation

Degree Name

Doctor of Juridical Science (SJD)

Abstract

Since the 1910s, Hollywood has measured audience preferences with rough industry-created methods. In the 1940s, scientific audience research led by George Gallup started to conduct film audience surveys with traditional statistical and psychological methods. However, the quantity, quality, and speed were limited. Things dramatically changed in the internet age. The prevalence of digital data increases the instantaneousness, convenience, width, and depth of collecting audience and content data. Advanced data and AI technologies have also allowed machines to provide filmmakers with ideas or even make human-like expressions. This brings new copyright challenges in the data-driven era.

Massive amounts of text and data are the premise of text and data mining (TDM), as well as the admission ticket to access machine learning technologies. Given the high and uncertain copyright violation risks in the data-driven creation process, whoever controls the copyrighted film materials can monopolize the data and AI technologies to create motion pictures in the data-driven era. Considering that copyright shall not be the gatekeeper to new technological uses that do not impair the original uses of copyrighted works in the existing markets, this study proposes to create a TDM and model training limitations or exceptions to copyrights and recommends the Singapore legislative model.

Motion pictures, as public entertainment media, have inherently limited creative choices. Identifying data-driven works’ human original expression components is also challenging. This study proposes establishing a voluntarily negotiated license institution backed up by a compulsory license to enable other filmmakers to reuse film materials in new motion pictures. The film material’s degree of human original authorship certified by film artists’ guilds shall be a crucial factor in deciding the compulsory license’s royalty rate and terms to encourage retaining human artists. This study argues that international and domestic policymakers should enjoy broad discretion to qualify data-driven work’s copyright protection because data-driven work is a new category of work. It would be too late to wait until ubiquitous data-driven works block human creative freedom and floods of data-driven work copyright litigations overwhelm the judicial systems.

Download

Share

COinS