The Legal Implications of Artificial Intelligence in Intellectual Property (IP)

Author: Kainaat Afreen

ABSTRACT
Artificial Intelligence (AI) is redefining the boundaries of creativity and innovation. Once limited to human authors and inventors, Intellectual Property (IP) law now faces the challenge of addressing works and inventions produced wholly or partly by machines. AI can generate music, art, literature, technical designs, and even propose scientific solutions, sometimes with minimal or no human intervention. Yet, the majority of IP laws worldwide, from copyright statutes to patent regimes, remain grounded in a human-centric model.

This mismatch between technological capability and legal recognition creates uncertainty. Fully autonomous AI outputs may fall into the public domain, discouraging investment, while AI-enabled infringement can occur at unprecedented speed and scale. Courts have started grappling with such issues, as seen in the UK and Australian DABUS patent cases, the US “Zarya of the Dawn” decision, and even analogies drawn from the “Monkey Selfie” litigation.

This article examines the implications of AI in copyright, patents, trademarks, and trade secrets, referencing statutory frameworks, international treaties, and judicial interpretations. It identifies the legal gaps, analyzes leading case law, and offers recommendations for legislative reform and international harmonization to ensure IP systems remain relevant and fair in the age of machine creativity.

TO THE POINT
AI’s role in creative and inventive activities has progressed from assistance to autonomy. This evolution confronts IP law with three pressing challenges:
1. Authorship and Inventorship – Can an AI be legally recognized as the originator of a work or invention? Under existing frameworks, the answer is generally “no,” with IP statutes interpreting “author” and “inventor” as natural persons.
2. Ownership and Accountability – If AI cannot hold rights, who does? Candidates include the programmer, AI owner, commissioning client, or operator providing prompts. This becomes more complex when multiple parties contribute to the AI’s design or deployment.
3. Infringement and Enforcement – AI can reproduce or modify protected works instantly and at massive scale. Identifying infringing content and holding the responsible party liable is complicated when the AI operates autonomously or is used across multiple jurisdictions.

Practical Concerns:
Businesses using AI for creative or technical output may fail to secure protection for works generated without significant human involvement.
Rights holders are at increased risk of having their works ingested into AI training datasets without permission, potentially leading to unlicensed derivative outputs.
Jurisdictional inconsistencies amplify uncertainty, making international enforcement difficult.
These issues show why IP law must evolve. Without change, innovators face disincentives to invest in AI technologies, rights holders struggle to safeguard works, and global markets experience uneven protections. Legislators, policymakers, and international bodies like WIPO have a critical role in ensuring IP law adapts to AI’s transformative potential.

USE OF LEGAL JARGON
Understanding AI’s IP challenges requires familiarity with key legal terms:
Originality Requirement – The minimum creative threshold a work must meet for copyright.
Inventive Step (Non-Obviousness) – A patent law criterion requiring that an invention not be obvious to a “person skilled in the art.”
Doctrine of Sweat of the Brow – A limited doctrine granting copyright based on effort or labor rather than creativity.
Non-Human Authorship Doctrine – The legal stance that only human beings can be authors or inventors.
Public Domain Encroachment – Expansion of private rights that reduces works freely available to the public.
Strict Liability in Infringement – Liability without proof of intent, relevant where AI autonomously generates infringing works.
Work for Hire – A rule assigning authorship to an employer or commissioning party, potentially applicable in AI-human collaborations.
Derivative Work – A work based on an existing one, central to disputes over AI outputs trained on copyrighted materials.
Moral Rights – Personal rights of authors, such as attribution, which raise unique questions when no human author exists.

THE PROOF
Statutory Context
Copyright:
United States – The U.S. Copyright Office refuses to register works lacking human authorship, as reiterated in the “Zarya of the Dawn” decision.
United Kingdom – Section 9(3) of the Copyright, Designs and Patents Act 1988 names as author the person making “arrangements necessary” for computer-generated works, though courts interpret this narrowly.
European Union – Originality is tied to the “author’s own intellectual creation,” which presumes human origin.

Patents:
U.S. Patent Act – Defines “inventor” as an “individual,” interpreted as a human being.
European Patent Convention – Article 81 requires naming a natural person.
Australia – In the DABUS cases, courts confirmed that inventorship is reserved for humans.

Trademarks:
AI aids in brand protection by identifying counterfeits and monitoring marketplaces, but cannot own marks.

Trade Secrets:
Laws like the Defend Trade Secrets Act (U.S.) and EU Trade Secrets Directive protect confidential algorithms and training datasets.

International Instruments
Berne Convention (Art. 2) – Lists protected works without AI-specific provisions.
TRIPS Agreement – Sets minimum IP protection standards, silent on AI.

Regulatory Trends
WIPO’s Conversation on IP and AI (2019–2023) highlights authorship, inventorship, infringement, and data use as priority areas for reform.

CASE LAWS
1. Thaler v. Comptroller-General of Patents, Designs and Trade Marks (UK, 2021)
Background: Dr. Stephen Thaler filed patent applications naming his AI system, DABUS, as the inventor. The UKIPO rejected the filings, stating that an inventor must be a natural person.
Court Decision: The Court of Appeal upheld the rejection. The majority reasoned that the Patents Act 1977 contemplates inventors as natural persons for accountability and rights enforcement purposes.
Significance: This ruling reinforced the position that AI cannot be named as an inventor, sparking debate on whether legislative reform is necessary to address AI-generated inventions.

2. Thaler v. Commissioner of Patents (Australia, 2022)
Background: Similar to the UK case, Thaler’s Australian patent filings listed DABUS as the inventor. A lower court initially accepted this, marking a global first.
Court Decision: On appeal, the Full Federal Court reversed the decision, holding that the Patents Act requires an inventor to be human.
Significance: The case illustrates judicial reluctance to deviate from human-centric inventorship without legislative change.

3. Naruto v. Slater (U.S., 2018)
Background: A macaque named Naruto captured images with a wildlife photographer’s camera, leading PETA to file a lawsuit asserting that the animal held the copyright.
Court Decision: The Ninth Circuit dismissed the claim, holding that non-humans cannot hold copyright under the U.S. Copyright Act.
Significance: Though involving an animal, this case is frequently cited in AI debates as an analogy against granting rights to non-human creators.

4. Zarya of the Dawn (U.S., 2023)
Background: A graphic novel was submitted for registration, containing text written by a human and images generated by the AI program Midjourney.
Court Decision: The U.S. Copyright Office granted protection for the human-authored text but refused protection for the AI-generated images.
Significance: This decision confirms that AI outputs without human creative control are not copyrightable in the U.S.

5. Feist Publications v. Rural Telephone Service Co. (U.S., 1991)
Background: Feist used Rural’s white pages listings in its own directory without permission. Rural sued for copyright infringement.
Court Decision: The U.S. Supreme Court held that factual compilations without creative selection or arrangement lack originality and are unprotected.
Significance: For AI, this suggests that factual datasets generated without creative human input may not qualify for protection.

CONCLUSION
AI is transforming creativity and invention, but IP law’s foundations remain human-focused. Existing laws prevent AI from being recognized as an author or inventor, placing entirely autonomous creations in a legal gray area. This preserves human accountability but risks discouraging AI-driven innovation.

Recommended Actions:
1. Amend statutes to define how AI contributions are treated, potentially via hybrid models recognizing human-AI collaboration.
2. Use WIPO to create harmonized international standards.
3. Develop technology-based enforcement measures like blockchain registries.
4. Adopt ethical safeguards to preserve the public domain and prevent monopolization of AI-generated culture.

Without coordinated reform, IP systems will remain fragmented and ill-equipped for the realities of AI.

FAQs
Q1: Can AI hold copyright?
No. Laws worldwide require a human author.
Q2: Can AI be an inventor?
No, under current statutes in major jurisdictions.
Q3: Is AI training on copyrighted data infringement?
It can be, depending on how the data is used and whether exceptions like fair use apply.
Q4: How is AI used in trademark enforcement?
By detecting counterfeit products and unauthorized uses of marks online.
Q5: Are there global rules for AI and IP?
Not yet, but WIPO is exploring potential frameworks.

Leave a Reply

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