Home » Blog » THALER V PERLMUTTER

THALER V PERLMUTTER

Authored By: Asokere Deborah

Lagos State University- Graduate

Case Title and Citation 

Thaler v Perlmutter, No. 1:22-cv-01564 (D.D.C. Aug. 18, 2023). 

Note: The defendant, Shira Perlmutter, was sued in her official capacity as the Register of  Copyrights, representing the United States Copyright Office (USCO). 

Court and Bench 

United States District Court for the District of Columbia (Howell J). 

Date of Judgment 

18 August 2023. 

Parties 

Plaintiff: Stephen Thaler 

Defendant: United States Copyright Office (USCO) represnted by Perlmutter. 

Facts of the Case 

Stephen Thaler, the plaintiff, developed an artificial intelligence system known as the  “Creativity Machine,” which autonomously generated a two-dimensional artwork titled ‘A  Recent Entrance to Paradise’. Thaler applied to the United States Copyright Office (USCO) for  registration of the work, identifying the AI system as the sole author and himself as the owner  by virtue of ownership of the machine.1 

The USCO refused Thaler’s application on the basis that copyright protection under United  States law is limited to works of human authorship. Thaler sought reconsideration, arguing that  denying protection to AI-generated works was inconsistent with the objectives of copyright  law.2 The refusal was affirmed on reconsideration, after which Thaler commenced an action  against the Register of Copyrights, Shira Perlmutter, in her official capacity before the District  Court. 

Issues Raised for Determination 

  1. Whether a work generated autonomously by an artificial intelligence system qualifies  for copyright protection under United States copyright law. 
  2. Whether an artificial intelligence system can be recognised as an “author” for the  purposes of copyright registration. 

Arguments of the Parties 

Plaintiff’s Arguments: 

The plaintiff, in person of Stephen Thaler, argued that the United States Copyright Act  (referred to as Title 17 of the United States Code (17 U.S.C)) not expressly limit authorship  to human beings and that denying protection to AI-generated works would undermine the  constitutional objective of promoting the progress of science and the useful arts.3 Thaler  contended that copyright law should evolve in response to technological advancement and  that refusing protection to AI-generated works would discourage innovation and investment  in artificial intelligence technologies. 

He further submitted that ownership and control of the AI system should justify recognition  of copyright in the resulting work, drawing an analogy between AI systems and traditional  tools used by human creators. Just as a camera or a musical instrument enables a human to  produce a work, AI functions as an advanced tool that, while capable of autonomous  operation, remains under the direction and guidance of its human operator. According to the  plaintiff, the mere fact that a work is produced autonomously by an AI should not, in itself, preclude copyright protection, especially where a human has actively initiated, supervised,  and shaped the creative process. He argued that the human’s role in setting the parameters,  selecting input data, and determining the creative direction constitutes sufficient intellectual  contribution to warrant recognition as the author. In this framework, the AI is viewed not as  an independent creator but as an instrument through which human creativity is expressed, and  denying protection would ignore the human agency that underpins the work’s conception and  execution. 

Defendant’s Arguments: 

The United States Copyright Office, represented by the Register of Copyrights, Shira Perlmutter, maintained that copyright law has consistently required human authorship. It  relied on statutory interpretation, legislative history, and longstanding judicial precedent  establishing that copyright subsists only in works created through human intellectual effort.4 

In contrast, the United States Copyright Office (USCO) argued that recognising AI systems  as authors would fundamentally alter the traditional structure of copyright law, which has  always been premised on human creativity and intellectual contribution. The USCO  emphasized that copyright is intended to reward the personal, original expression of human  authors, and extending authorship to non-human entities would exceed the statutory  framework and historical understanding of the law. It further contended that questions  regarding innovation, economic incentives, and the promotion of creative progress are policy  matters for Congress rather than the judiciary. Any expansion of copyright protection to  include AI-generated works, the USCO asserted, must be expressly authorised by legislation,  as such a move involves complex considerations about ownership, economic benefit, and  public access to creative works. The office cautioned that judicial recognition of AI  authorship could create legal uncertainty and disrupt the careful balance that copyright law  seeks to maintain between protecting creators and serving the public interest. 

Judgment 

The court upheld the decision of the United States Copyright Office and dismissed the  plaintiff’s claim in its entirety. It affirmed that copyright protection under the Copyright Act of 1976 is confined to works of human authorship and that a work generated autonomously  by an artificial intelligence system does not qualify for copyright registration under the  existing statutory framework. 

In reaching this conclusion, the court emphasised that although the Copyright Act does not  expressly define the term “author,” the concept has been consistently interpreted through  judicial precedent and administrative practice as requiring human creative involvement. The  court relied on the longstanding position of the United States Copyright Office, which has  repeatedly refused registration for works lacking human authorship, and noted that this  interpretation aligns with prior judicial decisions rejecting copyright claims by non-human  actors. 

The court further rejected the plaintiff’s argument that ownership or control of the artificial  intelligence system could substitute for authorship. It held that copyright law does not  recognise authorship arising merely from ownership of the means by which a work is  produced, particularly where the output is generated without human creative choice or  intellectual judgment. 

Importantly, the court underscored the limits of judicial authority in addressing technological  developments. It observed that extending copyright protection to AI-generated works would  constitute a significant policy decision with wide-ranging implications, a determination that  falls within the legislative competence of Congress rather than the judiciary. Until such  legislative reform occurs, the court maintained that the existing human-centred interpretation  of authorship must continue to govern copyright registration. 

Legal Reasoning and Ratio Decidendi 

The court’s reasoning was grounded in a historically consistent interpretation of authorship  under United States copyright law. Although the Copyright Act does not expressly define the  term “author,” judicial precedent has long associated authorship with human creative  activity.5 The court relied on earlier authorities rejecting copyright claims by non-human actors, reaffirming that copyright protection presupposes human intellectual effort, creative  choice, and intention. 

The court rejected the plaintiff’s argument that statutory silence permitted recognition of AI generated works. It held that silence could not be construed as legislative approval of non human authorship, particularly in light of settled judicial and administrative practice.6 The  court further emphasised that extending authorship to artificial intelligence systems would  amount to a significant policy shift better suited for legislative determination. 

With respect to ownership, the court clarified that copyright law does not recognise  authorship arising solely from ownership or control of the means of production. Unlike  human employees or agents, artificial intelligence systems lack legal capacity, intention, and  accountability.7 Accordingly, attributing authorship to the owner of an AI system without  evidence of human creative contribution would undermine the originality requirement. 

The ratio decidendi of the case was that copyright protection under United States law is  confined to works of human authorship, and that works generated autonomously by artificial  intelligence systems fall outside the scope of protection under the existing legal framework. 

Conclusion and Observations 

The decision in Thaler v Perlmutter reaffirms the human-centred foundation of copyright law  and provides authoritative clarification on the legal status of AI-generated works under  United States law. By insisting on human authorship as a prerequisite for protection, the court  preserved doctrinal coherence and avoided an expansive interpretation that could destabilise  existing copyright principles. 

At the same time, the judgment exposes a regulatory gap in the face of rapidly advancing  artificial intelligence technologies. As AI systems increasingly produce commercially  valuable creative outputs without direct human involvement, the absence of protection raises  questions about incentive structures, market regulation, and the appropriate scope of  intellectual property rights. 

Comparative Perspective: Implications for Nigeria 

The reasoning in Thaler v Perlmutter sets a particularly instructive precedent for Nigeria.  Nigerian copyright law, under the Copyright Act 2022, similarly presumes human authorship  and grounds protection in the exercise of human skill and effort. Like the United States  framework, it does not expressly contemplate works generated autonomously by artificial  intelligence systems. As AI adoption grows within Nigeria’s creative and technology sectors,  courts may soon be confronted with disputes analogous to Thaler. 

Absent legislative guidance, Nigerian courts are likely to adopt a conservative, human centred interpretation of authorship consistent with existing doctrine. While this approach  promotes legal certainty, it also highlights the need for proactive reform. Drawing from  comparative models such as the United Kingdom’s statutory treatment of computer-generated  works, Nigeria may need to clarify ownership rules, disclosure obligations, and the threshold  of human involvement required for protection. Without such reform, Nigerian copyright law  risks lagging behind technological realities and leaving significant categories of creative  output in a state of legal uncertainty.

Reference(S):

1 US Copyright Office, Copyright Registration Guidance: Works Containing AI-Generated  Material (2023)

2 Thaler v Perlmutter, No. 1:22-cv-01564 (D.D.C. Aug. 18, 2023) 

3 US Const art I, § 8, cl 8.

4 US Copyright Office, Copyright Registration Guidance: Works Containing AI-Generated  Material (2023)

5 Naruto v Slater 888 F 3d 418 (9th Cir 2018).

6 Thaler v Perlmutter, No. 1:22-cv-01564 (D.D.C. Aug. 18, 2023). 

7 Community for Creative Non-Violence v Reid 490 US 730 (1989)

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top