Reading Summary - AI Copyright EIPR-paper

The advancement of Artificial Intelligence (AI) in the 21st century has enabled machines to imitate human actions.
Copyright laws were traditionally established for humans, as seen with historical documents like the Statute of Anne (1710) and the Berne Convention (1886).
Key treaties defining copyright protection internationally include:
- Berne Convention (1886): Establishes that its member countries protect the rights of authors in their artistic works.
- TRIPS (1994): Trade-Related Aspects of Intellectual Property Rights addresses copyright issues.
- WCT (1996): World Intellectual Property Organization Copyright Treaty.

Article 1 of the Berne Convention suggests a synonymous relationship between natural persons and authors, implying that authors must be human beings.
This understanding is based on the traditional view that a creator is a human writing, painting, or composing.
Reflection on historical and ethical implications of treating AI-generated work in a human-centric framework.

The Next Rembrandt Project: Demonstrates the skill of machines in creating art.
UK Legal Recognition:
- Only UK law addresses “computer-generated works” through Section 9(3) of the Copyright, Design and Patents Act 1988 (CDPA).
- It assigns authorship to the individual or entity responsible for the arrangements needed to create the work, introducing complexity in attributing authorship when machines are involved.

China, a leader in AI technology, lacks a provision equivalent to the UK’s arranger provision but has made strides through landmark cases analyzing AI-generated works.
This paper aims to explore crucial cases from the People’s Republic of China (PRC) regarding copyright in AI-generated works, which have not been deeply examined in Western academic literature.

Examination of two landmark PRC cases on AI-generated works.
Discussion on the necessity of legal personhood for rights assignment associated with authorship and liabilities, promoting the work-made-for-hire (WMFH) doctrine.
Exploration of the originality criteria for AI-generated works, suggesting that a broader interpretation is needed for assessing the ingenuity between human creativity and machine creativity.
Comparison between CDPA provisions for defining authorship and ownership for AI-generated works, particularly in relation to the US and Chinese legal contexts.

Background: In Beijing Film Law Firm v. Baidu Netcom Technology Co., Ltd (2018), a Beijing law firm published an article without permission, which was subsequently replicated by Baidu.
Court Findings:
- Graphs: Considered non-copyrightable since derived from automatic software without natural person's creativity.
- Written Analysis: Not copyrightable as it is generated by non-human entities; neither the software developer nor user was deemed an author based on PRC Copyright Law that requires creation by natural persons.
- The written texts were copyrightable, affirming authorship belongs to the claimant as a legal entity.
Outcome: Baidu was liable for copyright infringement and required to pay damages and issue a public apology.

Background: In Shenzhen Tencent Computer System Co., Ltd v. Shanghai Yingxun Technology Co. (2019), Tencent's AI software Dreamwriter was used to generate articles, which were later reproduced by the defendant.
Court Findings:
- The court deemed the article copyrightable, highlighting the human involvement in stages such as data input and data selection, ensuring that the article encapsulated the creativity of the users.
Outcome: Tencent's copyright was affirmed; the court ruled the defendant liable for copyright infringement.

Current interpretations of copyright law necessitate that only “natural persons” qualify as authors, a practice mirrored in the US and EU.
The requirement for authorship being limited to human contributions raises barriers for AI-generated works, leading to debates over ownership and originality.

A rigid interpretation results in AI-generated works being labeled as “authorless,” thus unprotected by copyright.
The Baidu court indicated that while the AI-generated work showcased originality, it wasn’t a recognized “work” under Copyright Law as it wasn't created by a natural person.
The decisions emphasize the need for a legal framework that allows for the protection of AI-derived creativity.

Originality as a requirement for protection varies internationally and manifests through:
- Sweat of the Brow: Focuses on effort and skill without requiring extreme creativity (e.g., UK).
- Minimal Creativity: Requires a certain degree of intellectual creativity (e.g., US).
- PRC Version: Claims originality as “intellectual achievements.”

Baidu: The graphs were deemed not original due to their mechanical derivation.
Tencent: The court recognized that creative team involvement does create original work alongside the AI´s outputs, positioning human input as vital to generating unique articles.

Section 9(3) of the CDPA mandates that authorship be given to the person responsible for the arrangements that generate a work.
The unique situation arising from AI-generated works necessitates adaptation of the CDPA to address challenges posed by collaborative efforts between human programmers and AI.

Calls for potential adoption of CDPA-like provisions in Chinese law to provide clarity and recognize both AI developers and users as authors, thereby attributing copyright appropriately.
Contemplation on how the roles of AI developers and users intermingle with copyright law adaptation, particularly as AI technologies become further integrated into creative processes.

AI-generated works necessitate a reconsideration of traditional copyright frameworks to adapt to evolving technological landscapes.
While the court decisions affirm the creative potential of AI, they also underline the need for contemporary legal interpretations that provide adequate protections for all forms of creativity, paving the path for future discussions on AI's place in copyright law.