REASSESSING INVENTORSHIP IN THE AGE OF MACHINES
Abstract
Artificial Intelligence (AI) has transformed industries by enabling machines to create solutions once believed to be exclusively within the realm of human intelligence. From designing medical devices to developing complex algorithms, AI-driven systems increasingly contribute to innovation. However, this technological progress raises a fundamental legal question: Can an artificial intelligence system be recognised as an inventor under patent law? Traditional patent regimes were drafted with human inventors in mind, creating uncertainty regarding ownership, inventorship, and accountability for AI-generated inventions. This blog examines the legal challenges surrounding AI inventorship, with a particular focus on Indian patent law and global developments in the debate over patentability of AI-generated inventions.
Introduction
Innovation has historically been linked to human intellect. Patent systems across the world were developed on the assumption that inventions emerge from human creativity and ingenuity. However, the rapid advancement of Artificial Intelligence (AI) is beginning to challenge this foundational assumption.
AI systems are now capable of independently generating inventions without direct human intervention. A notable example is DABUS (Device for the Autonomous Bootstrapping of Unified Sentience), an AI system designed by Dr. Stephen Thaler, which allegedly created inventions independently. This development sparked a global legal debate on whether a machine can be legally recognised as an inventor.
As AI increasingly contributes to technological advancement, patent law faces a pressing challenge: must inventorship remain exclusively human, or should the legal framework evolve to accommodate machine-generated innovation?
Understanding Patentability and Inventorship
A patent grants exclusive rights to inventors over their inventions for a limited period, encouraging technological progress and innovation. Under the Patents Act, 1970, an invention must satisfy three essential requirements:
- Novelty – the invention must be new;
- Inventive Step – it must not be obvious to a skilled person; and
- Industrial Applicability – it should be capable of practical use.
However, patent law also emphasises the identity of the inventor. Section 2(y) of the Patents Act, 1970 defines a “true and first inventor” while explicitly excluding the “first importer” of an invention. Although the legislation does not expressly state that an inventor must be human, the statutory framework assumes human agency in inventorship.
This assumption becomes problematic when inventions are generated autonomously by AI systems.
The DABUS Controversy: Testing the Boundaries of Patent Law
The most significant development in AI inventorship emerged through the DABUS litigation, where patent applications listed an AI system as the sole inventor.
United Kingdom
In Thaler v Comptroller-General of Patents, Designs and Trade Marks (2023), the UK Supreme Court rejected patent applications naming DABUS as the inventor. The Court held that under existing UK patent law, an inventor must be a natural person, thereby excluding AI systems from legal recognition.
United States
Similarly, in Thaler v Vidal (2022), U.S. courts concluded that only human beings qualify as inventors under the Patent Act. The judgment stressed that statutory language consistently refers to inventors as individuals.
Australia
In Thaler v Commissioner of Patents [2021] FCA 879, the Federal Court of Australia initially held that an AI system could, in principle, be named as an inventor under Australian patent law. However, this position was overturned by the Full Court of the Federal Court in Commissioner of Patents v Thaler [2022] FCAFC 62, which ruled that the Patents Act contemplates only a natural person as an inventor. As a result, Australia ultimately aligned itself with the United Kingdom and the United States in refusing recognition of AI as an inventor.
South Africa
South Africa, however, granted a patent listing DABUS as an inventor, making it one of the few jurisdictions willing to experiment with AI-generated patent recognition.
The inconsistency in global responses reflects the uncertainty surrounding AI inventorship and highlights the absence of a universally accepted legal framework.
The Indian Position on AI Inventorship
India presently lacks specific legal provisions addressing AI-generated inventions. The Patents Act, 1970, drafted long before AI innovation became relevant, does not contemplate autonomous machine inventorship.
Indian patent law strongly associates inventorship with legal personality and accountability. Since AI lacks legal personhood, practical concerns emerge:
- Ownership Issues: If an AI system independently develops an invention, who owns the patent rights?
Potential claimants may include: the programmer of the AI; the company deploying or owning the AI system; or the individual directing and supervising the inventive process.
While these competing claims remain unresolved in cases involving fully autonomous AI, patent offices across major jurisdictions have generally adopted a practical distinction between AI-generated and AI-assisted inventions. Where AI merely assists a human inventor by analysing data, generating alternatives, or accelerating research, the human who exercises meaningful intellectual contribution and control over the inventive process is ordinarily recognised as the inventor.
This approach has been reflected in the practices of patent offices such as the United States Patent and Trademark Office (USPTO), the European Patent Office (EPO), and the United Kingdom Intellectual Property Office (UKIPO), all of which continue to require a natural person to be identified as the inventor. As AI increasingly functions as a collaborative tool rather than an independent creator, determining the extent of human contribution is likely to become the central question in future patent disputes. - Accountability and Liability: Patent systems rely on identifiable inventors for legal accountability. AI systems cannot be held legally responsible for misrepresentation, fraud, or patent disputes.
- Innovation Policy Concern: Denying patent protection to AI-generated inventions may discourage investment in AI-driven research. At the same time, unrestricted recognition of machine inventorship could undermine the human-centric philosophy underlying intellectual property law. India must therefore strike a balance between encouraging innovation and preserving legal certainty.
AI-Assisted Inventions: The Current Practical Position
While litigation concerning fully autonomous AI inventors has attracted considerable attention, most patent applications today involve AI-assisted inventions rather than inventions created entirely by AI. Patent offices across major jurisdictions continue to require that a natural person be identified as the inventor. Where AI functions as a sophisticated research tool and a human exercises intellectual control over the inventive process, patent protection generally remains available. This practical distinction enables innovation to continue without fundamentally altering the traditional concept of inventorship.
For Indian applicants, this distinction is particularly significant, as innovations developed with the assistance of AI are becoming increasingly common in research and industry. Clarifying the standard for human inventorship through administrative guidance would provide greater certainty while preserving the existing framework of the Patents Act.
Should AI Be Recognised as an Inventor?
The debate over AI inventorship remains deeply contested.
Arguments Supporting Recognition of AI Inventors
- AI increasingly contributes meaningfully to scientific innovation.
- Patent denial may discourage technological investment.
- Existing frameworks risk becoming outdated in the face of autonomous systems.
Arguments Against Recognition
- AI lacks legal personality and moral accountability.
- Patent law traditionally rewards human ingenuity.
- Recognising AI inventors may complicate ownership and enforcement mechanisms.
A practical middle path could involve recognising human controllers or developers as inventors, while acknowledging AI’s role in the inventive process. Such an approach may preserve accountability without discouraging innovation.
Conclusion
Artificial Intelligence has fundamentally altered the landscape of innovation, compelling legal systems to reconsider long-standing assumptions surrounding inventorship. While AI-generated inventions promise scientific advancement, they also expose significant gaps within existing patent frameworks.
The DABUS litigation demonstrates that most jurisdictions remain reluctant to recognise machines as inventors, preferring to preserve human-centric patent systems. India, too, currently lacks legislative clarity regarding AI inventorship under the Patents Act, 1970.
Rather than immediately recognising AI systems as inventors, India should consider issuing administrative guidelines that clarify the treatment of AI-assisted inventions while retaining the requirement that inventorship be attributed to a natural person exercising meaningful intellectual contribution. Such an incremental approach would provide certainty to innovators without disrupting the foundational principles of patent law. As AI capabilities continue to evolve, Parliament may subsequently evaluate whether broader legislative reforms are necessary.
Nevertheless, as AI systems increasingly contribute to technological breakthroughs, maintaining rigid legal definitions may prove unsustainable. The future of patent law may not necessarily involve recognising machines as inventors, but it will certainly require laws capable of addressing inventions created in partnership between humans and intelligent machines.
References
The Patents Act, 1970 (India).
Thaler v Comptroller-General of Patents, Designs and Trade Marks [2023] UKSC 49.
Thaler v Vidal, 43 F.4th 1207 (Fed. Cir. 2022).
Thaler v Commissioner of Patents [2021] FCA 879
Commissioner of Patents v Thaler [2022] FCAFC 62
World Intellectual Property Organization (WIPO), WIPO Technology Trends: Artificial Intelligence (2019), available at: https://www.wipo.int
European Patent Office, Artificial Intelligence and Patentability Guidelines (2024), available at: https://www.epo.org
Abbott, R. (2020). The Reasonable Robot: Artificial Intelligence and the Law. Cambridge University Press.
Samuelson, P. (2021). “Implications of Artificial Intelligence for Intellectual Property Law”, Harvard Journal of Law & Technology.