AI, Authorship, Copyright, and Human Originality

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper is interesting and can be published. It attempts to present a new framework to analyse the originality requirement in copyright law, without which a work cannot (under current law) be protected under the Berne Convention.   

There are, however, a few doctrinal errors and shortcuts that weaken the analysis and must be addressed before publication.

First, it is not a given (I have checked this with two UK based copyright law professors) that computer generated works in UK law can be protected if devoid of human originality.  If Berne was predigital, as the authors note, that amendment to UK law was pre AI. 

Second, the Naruto v Slater case cannot be cited for the proposition advanced by the authors.  the case stands for the proposition that a nonhuman does not have standing to sue. The correct case to cite is Thaler v Perlmutter (DC Circuit). 

Third, the statement that image and likeness rights in the US is mostly state law is correct, but that does not mean it is weak. It is also incorrect or at least potentially misleading to say that it is essentially "unfair competition" law. Then it may be worth pointing out that a bill is pending in Congress. 

Fourth, the insistence on moral rights may be partially misplaced: It may well be that non copyright legislation is more suitable to protect name, image and likeness (NIL) rights.                

Fifth, the statement that Berne is "silent on hybrids" is also potentially misleading. A statute or treaty can be interpreted to apply to new situations. If the authors mean that there is no *explicit* rule in Berne concerning AI (obviously, since the last update was in 1967 except for the addition of an Appendix in 1971), they should say so.

Sixth and finally, the LAOIN case is under appeal and in the meantime, it looks like a Munich court will rule somewhat differently in the GEMA case (expected, I believe on 11 November). Emphasis on the first instance decision is LAION should thus be limited and, if available, the GEMA case should be added to the analytical mix.

 

 

Author Response

Please see uploaded PDF document

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dear authors, 

I appreciate your effort to deal with this hot topic and the challenges that AI brings. However I cannot share with you the use of the proposed MATH-COPE matrix.  I miss the analysis of other rights, especially the exploitation rights because they are required to use any work and they are the ones that can be commercialised.

I can agree with you about your proposals on section seven but I miss some reflexion abut the current situations. For instance on point 3 (Training data license) there is no mention about reproduction that is already a protected  exploitation right. Moreover some aspects as voice, style should be also covered by personal rights that are not  mentioned. What about the GDPR? I see also a lack of explanation on the proposed exceptions that seem similar than the existing ones. Moreover exceptions apply to exploitation rights. I don't see either any discussion about the term of the proposed protections

I also miss a look at the European TDM exception. The text only mentions the UK exception.

I am sorry but I don't see this document suitable for being published in this journal

 

Author Response

Please see uploaded PDF document

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors Some suggestions:​ 1) There are layout errors in the document, such as the stray text "92" visible at the top of Page 3. Please check the manuscript for clean formatting.​ 2) Figure 1 provides limited informational value. I suggest either improving it to add more insight or removing it.​ 3) Many items currently labeled as "Figures" are essentially text-based matrices (e.g., Figures 2, 3, 4, and 5). These should be renamed as "Tables".​ 4) Certain arguments rely too heavily on a single source. For instance, Section 5 depends almost entirely on one reference. The authors should broaden the bibliography and include additional sources to support these claims and avoid over-reliance on a single text.

Author Response

Please see uploaded PDF document

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Thanks for improving the text

Author Response

Dear Reviewer,

Thank you for your positive review of our revised manuscript.

Best regards.