AI-Generated Innovation and Indian Patent Law: Challenges, Doctrinal Gaps, and Pathways Forward

 AI-Generated Innovation and Indian Patent Law: Challenges, Doctrinal Gaps, and Pathways Forward

I. Introduction

India’s technological rise in areas such as AI, biotechnology, and digital platforms is shifting its innovation ecosystem. With thriving hubs in Bengaluru, Hyderabad, and Delhi-NCR, AI systems are increasingly being used to generate designs, discover novel compounds, and optimize industrial processes. Traditionally, Indian patent law is built on human ingenuity—rooted in the Patents Act 1970 and aligned with TRIPS principles. Yet AI challenges fundamental doctrines: What constitutes inventorship? Who owns AI-derived inventions? This paper explores whether the existing legal framework can sustain the accelerating pace of AI innovation or if systemic reform is imperative.

II. Methodological Approach

This study employs a doctrinal-analytical method, focusing primarily on Indian statutes (the Patents Act, 1970; Patent Rules, 2003), judicial practice, and administrative procedures. Comparative jurisdictional references—including the UK’s DABUS case and Singapore’s emerging AI-IP reports—offer guiding insights, but Indian policy, market data, and public interest considerations remain central. We draw on:

• Government and WIPO data on AI patent filings;

• Filings involving AI systems in Indian Patent Office databases over the past three years;

• Doctrinal breakdowns of Sections 6 (inventorship), 10 (disclosure), and 3(k) (programs excluded from patentability);

• A review of Indian judicial trends around AI, such as the Ferid Allani doctrine;

• Industry interviews with patent professionals in Bangalore and Noida (quoted anonymously).

III. India’s AI Innovation Landscape

Since the 2018 NITI Aayog AI strategy, India has launched initiatives like AI4Bharat, CHiMERA hospital data projects, and AI-fusion smart agriculture pilot schemes.¹ Investments exceed USD 3 billion annually, with startups filing patent applications in areas like machine learning-assisted drug formulation.² WIPO’s 2023 data shows AI-related filings in India rose nearly 350% from 2017 to 2022.³ These trends highlight AI’s emerging role in co-developing IP—but current laws don’t accommodate non-human creativity.

IV. Doctrinating Inventorship and AI

A. Statutory Reflection

The Patents Act defines “inventor” under Sections 2(s) and 2(y), in combination with “true and first inventor” in Section 6, and refers to “person”—implicitly a natural individual.⁴ Applicant declaration forms and Rules 6, 6A² assume a human creator. As fielded in practice, systems simply do not allow AI names.

AI-generated filings (e.g., for process optimization systems) invariably list programmers or research supervisors as inventors. The IPO responds requesting clarification—calling out absence of human inventive conception or technical disclosure. No AI-named patent has entered the Indian system to date.

B. Disclosure Challenges

Section 10 mandates an enablement of “sufficient description” including the “best method known.” AI’s reasoning—particularly deep learning or neural models—often emulates reasoning by pattern, not explainable logic. Patent office objections center around missing bitwise pseudocode, parameter settings, or dataset transparency.⁵

A 2022 provisional application by an IIT–Delhi researcher demonstrated this. The form included sample code, architecture diagrams, training set characteristics—enough to comply and progress, but subsequent Office actions flagged lack of explainability on design heuristics.⁶ This represents a shifting compliance barrier for advanced AI.

V. Ferid Allani and the Technical Effect Precedent

In the landmark Ferid Allani v. UoI decision, the IPAB accepted computer-implemented inventions demonstrating technical effect (e.g., improved data security protocols implemented with algorithmic innovation) despite being understood partly as computer programs.⁷ The ruling empowers technical effect distinctions—useful for AI inventions tackling cyber security, IoT automation, and sensor-driven robotics. Yet it leaves unresolved whether logic discovered by AI independently qualifies.

VI. DABUS and India: Comparative Reference

Though India has yet to witness formal judicial examination of AI inventorship, the IPO’s rejection of a DABUS–based application in 2021 signals alignment with UK and EPO doctrine. Analysts argue India could follow the Common Law route, unless legislative intervention specifies otherwise. Reform proposals include expanding Section 6 to accommodate AI-assisted inventions if a “natural person” plays a substantial role in conception, or allowing “AI tool” primacy with human entitlement.

VII. AI Patent Examination Practices in India

Recent Patent Examination Guidelines (March 2024) reflect cautious progress—explicitly mentioning algorithm-generated outputs are considered if technical effects are demonstrated.⁸ Examiners rely heavily on comparative examples: noise-reduction filters for MRI scans, air traffic optimization modules, and AI grid computing solutions. Interviews with examiners reveal AI systems are treated as part of the inventive environment—yet all filings list human originators.

Dr. S., Deputy Controller at Chennai IPO, explained: **“We evaluate AI contributions the same as any other technical invention. The missing piece remains the attribution.”**⁹

VIII. Ownership, Entitlement, and Entanglement

Indian law attaches ownership to the inventor’s rights and any assignee in writing.¹⁰ Where AI is unacknowledged or named, entitlement rests with developers or employers. This preserves commercial viability but sidelines the philosophical question of AI’s role in invention creation.

Without reform, ownership may remain legally coherent but distort moral attribution—raising questions about who merited innovation recognition and public trust in equitable IP governance.

IX. Public Interest, Technology Equity, and AI Patents

Patent systems in India prioritize social development. Recent legislation (e.g., narrow opposition grounds, Section 3(d)) is used to ensure generic drug access.¹¹ AI may generate breakthroughs in generative drug algorithms or precision diagnostics. Allowing AI-indicated patents with strong disclosure may accelerate medical innovation—but also risk monopolistic barriers unless public-interest safeguards are implemented.

The 2024 amendment of insertion of compulsory license thresholds for pharmaceuticals, plastics, or engineered crops could apply equally to AI patents.¹²

A. Section 6 Revisions

Re-casting records to allow claims of inventive contribution by AI, accompanied by specifying a lead human inventor—modelled on Australia’s initial DABUS decision.

B. Regulatory Examination Framework

Detailed IPO guidelines that outline AI attribution, technical effect demonstration, and disclosure formats (e.g., repository of model code, datasets).

C. Hybrid Ownership

Separating inventor from entitlement: naming the AI-hybrid system as creator (for recognition) while legal rights go to the sponsoring company or human operator.

D. Public-Interest Licensing

Automatic licensing triggers for innovation in essential domains created via AI—balancing investment and access.

X. Concluding Synthesis

India stands at a critical juncture, balancing patent law’s role as an innovation enabler with technological complexity and public interest. The current human-centric doctrine limits AI’s full creative expression but provides legal clarity. Structural reform—calibrated to India’s developmental ethos—could preserve exclusivity while encouraging continued technological progress. In doing so, India would not merely keep pace globally, but lead reform in equitable technology governance.

Closing Credit 

Author - Bushra 

"The views expressed are personal. This article is intended for educational purposes and public discourse. Feedback and constructive criticism are welcome!"

---

OSCOLA Footnotes

1 DPIIT, National Strategy on Artificial Intelligence (Government of India, 2018).
2 NASSCOM–Zinnov, "Indian AI Start-up Ecosystem" (2023).
3 WIPO, Technology Trends 2023: Artificial Intelligence (WIPO 2023).
4 ibid.
5 Patents Act 1970 (India) s 10.
6 IPO Office Action, Application No. ________ (2022).
7 Ferid Allani v Union of India (Controller General of Patents) 2022 SCC OnLine Del 1224.
8 IPO Examination Guidelines (March 2024): see Chapter on AI and Machine Learning Applications.
9 Interview with Dr S. (Deputy Controller, Indian Patent Office Chennai), February 2025.
10 Patents Act 1970 (India) ss 6, 64.
11 Patents (Amendment) Act 2023 (India).
12 ibid s 92 (Compulsory licenses in public health).