“By far, the greatest danger of artificial intelligence is that people conclude too early that they understand it.” – Eliezer Yudkowsky

While Artificial Intelligence (AI) was initially introduced as a tool to help solve human problems by replicating human intelligence and combining it with a computer’s software, nowadays it is used for a myriad of purposes, from managing our calendars and schedules, generating data with human-like creativity, to being our personal therapist. It has revolutionized the world and people’s life, especially in the fields of art, pharmaceuticals and so on. As much as it seems useful in terms of easing our burden of menial tasks, or sometimes even doing those tasks in place of us, it has posed its own set of complications and problems, specifically in the industries of creativity, pharmaceuticals and industrial design.

Traditionally, Intellectual property laws and frameworks across the world have been written and framed on the basic assumption that only human beings can work and create inventions. Copyrights, patents, industrial designs and so forth have been credited to human work for their creativity, ingenuity, judgement and accountability. For these very reasons they have been awarded protection under IPR laws.

But since the arrival of Artificial Intelligence, there is a dam of complications and confusions released for the law makers (especially) to deal with. Firstly, with respect to Ownership- If AI generates something new or unique, who gets to legally own it- the AI, the programmer, the company deploying the AI, or do all of them have to share ownership? Secondly with regard to Accountability- AI has no legal recognition or personality of its own, so it cannot be held liable for plagiarism, infringement, or any kind of harm caused by its outputs. In such cases, who can be held responsible or accountable is a question mark. Lastly, with respect to recognition and fairness- denying AI inventorship might mean that the true source of the inventive step goes unacknowledged and unrecognized, but granting it such recognition brings the risk of undermining human-centered intellectual property systems. On top of all this, the laws regarding intellectual property rights are vague and unspecific about the inclusion or exclusion of AI and patent rights. This has a big risk of people exploiting this loophole for their personal and capitalistic gain.

Indian Frameworks on Patent Rights

Patent rights are governed in India under the Indian Patents Act, 1970. According to this act, a patent is essentially a simple right to the inventor in order to protect the rights of the inventor over their invention. Essentially, it means that only the inventor (patent holder) has the right to use, sell, make or license the product or invention for a specific fixed period of time. It also contains provisions that helps balance rights of the inventor along with public interest in case of certain inventions. For instance, inventions in pharmaceuticals that might be useful to the public should not be sold at exorbitant rates and essentially prevent accessibility for the downtrodden and lower class of the society.

The following are some of the sections from the Indian Patents Act, 1970 that discuss about patent rights and the eligibility criteria to obtain a patent:

Section 2(y): “True and First Inventor” Defined

This section defines the true and first inventor to be a person who does not comprise:

• The first importer of an invention into India

• An individual to whom an invention is first disclosed from outside India

Section 2(ab): Meaning of “Assignee”

The Act equates an assignee to include:

• An assignee of the assignee

• The legal representative of a deceased assignee

• References to assignees include assignees of legal representatives

Sections Regulating Assignment and Transfer of Patent Rights:

Section 6: Parties who can Apply for Patents

This initial section defines who is able to apply for and eventually own a patent, and they are as follows:

• Any individual who asserts to be the first and true inventor of the invention

• Any individual as the assignee of the first and true inventor in regard to the right to make such application

• Legal representative of any person who at the time of death was entitled to make such application

Section 6(2) lays down that the applications may be made alone or jointly with any other individual.

Section 68: Assignments Should Be Only in Writing

This is an important section that lays down the formal requirements of transfer of patent ownership.

Essential Requirements are as follows:

• Assignment should be in writing

• The contract should be properly executed

• Should contain all conditions and terms regulating rights and obligations

NOTE: Post-2005 Amendment Effect: Registration not required for validity – assignments are valid from date of execution, although registration is still recommended.

Section 69: Registration of Assignments and Transmissions

This section provides rules for formal registration, and they are as follows:

• Where any person becomes entitled to a patent by assignment, he shall apply in writing for registration of title

• The Controller will not receive an unregistered agreement as evidence for establishing patent title, except on recorded reasons

Section 70: Patent Owner’s Power to Deal with Patent

This section confers universal rights to patent owners, which are as follows:

Powers Include:

• Assign patent rights

• Grant licenses under the patent

• Otherwise deal with the patent

• Give effective receipts for consideration received

Important Limitation: Subject to co-ownership provisions and any registered third-party rights.

Sections Defining Rights of Patent Owners

Section 48: Patentees’ Rights

This section provides exclusive rights to owners of patents:

• Right to manufacture, use, exercise, sell or distribute the patented article

• Right to prevent others from manufacturing, using, offering for sale, selling, or importing without authorization

Effective for 20 years from date of filing

Section 50: Co-owners of Patents’ Rights

This detailed section regulates co-ownership of patents:

• Equal undivided shares unless agreement otherwise

• Each co-owner has a right to exercise patent rights independently for own account without being answerable to others

• Restrictions on licensing and assignment: Cannot license or assign shares without the authorization of other co-owners

• Purchaser protection: Buyers from one co-owner may deal with patented articles as if bought from sole patentee

• General property law application: Principles of movable property ownership apply to patents

DABUS Case and AI Inventorship

The DABUS case is a prominent and landmark case in the matter of whether or not Artificial Intelligence can be recognized as an inventor under patent laws across the globe. Developed by Dr. Stephen Thaler, DABUS stands for “Device for the Autonomous Bootstrapping of Unified Sentience”, and the primary purpose for its creation was for generating new ideas and inventions with little to no human intervention. Dr. Thaler attempted to file patent applications in various countries like UK, USA, EU, Australia, South Africa, and more, where he listed DABUS, ie- the AI system, as the inventor instead of himself. The following were his arguments:

• DABUS created the inventions autonomously.
• Since inventorship should go to the “actual creator,” the AI should be recognized as the inventor.
• Thaler claimed he should be the patent owner because he owned the AI.

In all major jurisdictions except for a brief Australian ruling, courts and patent authorities rejected AI inventorship and related ownership, holding that inventors must be natural persons. Rulings by some of the countries are:

1. United Kingdom (UK)

• The UK Supreme Court held unanimously in 2023 that an AI such as DABUS cannot be specified as an inventor under the Patents Act 1977.

• The Court asserted that only a “natural person” may be an inventor, and Dr Thaler (DABUS’s owner/creator) was not automatically entitled to the inventions by virtue of being the owner of DABUS.

• Patent applications naming DABUS as inventor were found to be invalid and withdrawn, confirming that patent rights and inventorship cannot accrue in an AI or be assigned from an AI to its owner.

1. European Union (EU)

• The European Patent Office (EPO) ruled that the “inventor” named in a patent application should be a human being, not an AI or legal person.

• The Legal Board of Appeal of the EPO reaffirmed that DABUS, being an AI, is not able to be the inventor under Article 81 (indicating inventor) and Article 60 (right to a European patent) of the European Patent Convention.

• Any European patent right has to come from or go through a human inventor, so DABUS “ownership” or inventorship was clearly dismissed.

1. United States (USA)

• The United States Patent and Trademark Office (USPTO), federal courts, and eventually the U.S. Supreme Court all concluded that only human beings are inventors under U.S. law.

• The Federal Circuit construed the patent statute’s use of “individual” to mean a natural person, so DABUS could not be an inventor.

• Dr Thaler’s contention that patent rights to shift from DABUS to him as the proprietor was dismissed, entirely slamming shut the door to both ownership and inventorship by AI.

1. Australia

• The Federal Court of Australia (2021) initially specifically held an AI (DABUS) could be an inventor, but this was immediately overruled by the Full Federal Court in 2022.

• The appeal court upheld that the Patents Act mandates an inventor to be a human being and that an AI cannot own rights or inventions-though there was acknowledgment of conceptual distinctions between inventorship and ownership.

• Presently, Australia stands with other key jurisdictions in excluding AI inventorship and ownership.

Policy and Ethical issues in AI Ownership

1. Policy Concerns and Reactions

India’s policy discussions have emphasized the necessity for reform, with the Parliament committee suggesting revised categories to safeguard AI and inventions related to AI, and with proposals to shift to an “innovation oversight” framework, where ownership is granted to the human who exercises intellectual control (operator or developer).

In contrast to EU or US, where the emphasis is placed on the function of a “significant human contribution” or technological effect, Indian law has no formal provisions, creating greater legal uncertainty and lower incentives for solid AI-based R&D and commercialization.

There are fears related to national competitiveness, as not doing this may lead India behind in international AI innovation and miss out on investment in advanced technology.

• Ethical Concerns: Ownership, Responsibility, and Attribution
• Ownership: The ethical dilemma is who should own patents over AI-created inventions-the developer, owner, operator, or trainer of the data for the AI. Blanket ownership by a single group might not be fair, since AI systems can run with high autonomy and collective inputs.

1. Responsibility: The “black box” character of AI makes responsibility for invention and for possible misuse or harm difficult to attribute. In the absence of overt human inventorship, liability and legal responsibility can become diffuse, with enforcement and consumer protection harder to implement.

1. Credit: Ethical credit for innovation should accurately account for human creativity and the contribution of AI. Excessive credit to the human operator can minimize the role of the AI, but attributing rights to AI subverts the system’s purpose to incentivize human creativity.

Suggested Legal Reforms

1. Revise the Patents Act to Directly Deal with AI Inventions

• Implement statutory definitions of “AI-assisted” and “AI-generated” inventions, making it clear that patent ownership should finally rest in a natural or legal person, yet acknowledging the contribution or tool role of AI.

• Envision a model where AI can be named as a co-inventor or credited for its contribution, without conferring legal personhood to AI, as discussed in EU and US academe proposals.

• Use the “Innovation Oversight” or “Substantial Human Contribution” Model

• Make patents of AI-generated inventions conditional upon a human showing significant intellectual contribution or oversight, as in USPTO’s guidelines and “Pannu factors”.

• Create a legislative three-step test to determine: (1) the level of human intervention; (2) the worth of the intellectual contribution made by the human; and (3) whether the individual should be the owner of the patent.

• Provide Flexibility for Joint and Collaborative Ownership

• Provide allowance for joint inventorship where contributors include AI designers, trainers, and operators, and make legal rights and obligations explicit between stakeholders.

• Modernize provisions related to co-ownership to allow easy assignment, licensing, and commercialization of innovations involving AI.

• Enlarge Definitions of Patentable Subject Matter

• Amendment of Section 3(k) so that any technical advances made through AI are not precluded merely because they use algorithms, if they have concrete technical consequences or introduce novel solutions to technical challenges.

• Issue new Computer-Related Inventions (CRI) guidelines with clear reference to subject matter related to AI.

• Enhance Ownership, Disclosure, and Accountability Requirements

• Impose the full disclosure of the extent of AI contribution under the specification for AI-based patent applications, to ensure transparency and clarity regarding inventorship and inventive step.

• Institute registries to duly acknowledge substantive AI contributions and avoid disputes over “hidden” AI inventorship.

• International Alignment and Cooperation

Collaborate with international IP organizations and implement best practices to enable Indian innovators to protect and enforce their AI inventions internationally.

In conclusion, the issue of acknowledging AI as an inventor under the Indian patent law system demonstrates severe gaps between established legal principles and modern technological development. Indian patent law, based on human-oriented conceptions of inventorship and ownership, does not yet have specific provisions for accepting inventions created independently by AI systems.

This legal void generates doubt over who rightfully owns patent rights-whether the AI developer, operator, or some other party-thus possibly hampering incentives for innovation.

Finally, determining the difficult questions of ownership, accountability, and credit in inventions created by AI is critical for India to have a healthy and visionary patent system. It is critical to adopt flexible and subtle legal structures that appreciate the cooperative nature of human-AI innovation while protecting the integrity of the Indian intellectual property regime in the machine age.

---

Author: Laasya Priya Venkatesan

---