The emergence of Artificial Intelligence (AI) has significantly transformed various sectors, notably in technology and education. The integration of AI into everyday applications has led to enhanced efficiencies and novel capabilities. In the legal domain, AI’s role is increasingly critical, particularly in patent law, where the question of inventorship and patentability of AI-generated inventions is under scrutiny. The Indian Patent Act, specifically Section 3(k), asserts that mathematical methods, business methods, and algorithms are not patentable. This limitation raises concerns about the ability to protect innovations developed through AI, which could hinder technological advancement and economic growth. The discourse surrounding AI also emphasizes the need for adaptive legal frameworks that can accommodate rapid technological changes. For instance, the Delhi High Court in the case of Microsoft Technology Licensing, LLC v. The Assistant Controller of Patents and Design¹ highlighted the necessity for clear guidelines to navigate the patentability of computer programs and AI-related inventions. Similarly, the case of OPENTV INC v. The Controller of Patents and Designs² pointed out the obstacles faced in recognizing AI as a legitimate inventor under current laws, advocating for a review of existing provisions to foster innovation.

Keywords: Artificial Intelligence, Section 3(k), Delhi High Court.

The emergence of Artificial Intelligence (AI) has significantly transformed various facets of human life.³ This technology, which is less than 60 years old, gained its name during a pivotal conference in the mid-20th century when the first digital computers began to emerge in academic settings. The attendees were primarily mathematicians and computer scientists, many of whom aimed to develop algorithms that could be executed on these machines. This gathering fostered a sense of optimism, bolstered by early achievements in AI, which led to exaggerated predictions about its potential. The prevailing belief was that if computers could tackle complex problems that humans struggle with, like proving mathematical theorems, they should also be able to address simpler issues on our behalf. However, this assumption proved flawed. It became evident that challenges difficult for humans could sometimes be straightforward for computers, and vice versa. This discrepancy is not surprising, considering that computers operate using mathematical logic, allowing them to excel in logical problems where humans might falter.⁴ The trajectory of artificial intelligence (AI) research has evolved significantly over the past six decades. Its initial phase, marked by the Dartmouth Conference5, centered on the concept of General Problem Solving (GPS) (Newell & Simon, 1971). This approach posited that any problem translatable into code—be it proving mathematical theorems, playing chess, or calculating the shortest route between cities—was solvable. This involved representing knowledge in a computer-compatible format and systematically searching through possible solutions. For instance, in chess, the program would contain symbolic representations of the board, pieces, potential moves, and optimal moves based on heuristics derived from past tournaments. During gameplay, the search algorithm would identify the most advantageous move. However, the GPS approach encountered limitations as the complexity of problems increased, due to the exponential growth of search combinations. This challenge is particularly relevant in India today, where AI is being applied to complex problems like optimizing logistics for e-commerce companies facing massive demand and diverse geographical challenges, or in agricultural planning where numerous variables like soil type, weather patterns, and crop yields need to be considered.

Report and Projects

Consequently, the second phase of AI research shifted focus to optimizing search strategies— reducing the search space and improving knowledge representation. This era witnessed breakthroughs like Shrdlu (Winograd, 1972), a natural language understanding program. However, the field faced a setback with the Lighthill Report (1973),6 which cast doubt on the practical applications of AI. Similar concerns arose in the US and elsewhere. Despite this, Japan’s ambitious Fifth Generation Computer Systems (FGCS) project (1982) reinvigorated AI research globally. This comprehensive project encompassed both hardware and software development, including intelligent software environments and fifth-generation parallel processing. This project served as a catalyst for renewed interest in AI, prompting initiatives like the UK’s ALVEY project, a collaborative effort between government, industry, and academia, which explored the feasibility of Intelligent Knowledge-Based Systems (IKBS), also known as expert systems, across numerous demonstration projects. India could benefit immensely from such targeted projects today, particularly in areas like healthcare diagnostics in remote areas where access to specialists is limited, or in education where personalized learning platforms powered by AI can address the diverse needs of students. This led to the third phase of AI research, which emphasized IKBS. Unlike the universal knowledge approach of GPS, IKBS relied on domain-specific knowledge. For example, an IKBS for diagnosing infectious diseases would incorporate expert medical knowledge. This knowledge, often derived from human experts or other sources, was typically formalized as rules. A typical rule from the medical expert system MYCIN (Shortliffe et al 1975) might be: IF (patient has fever AND patient has cough) THEN (patient may have pneumonia). The collection of rules and facts constituted the knowledge base, which an inference engine used to derive conclusions. IKBS achieved significant success, with systems like R1, MYCIN, and Prospector finding commercial applications (Darlington, 2000). These systems highlight the potential for AI in India’s rapidly growing healthcare sector, where AI-powered diagnostic tools can assist healthcare workers in making faster and more accurate diagnoses, especially in underserved rural communities.⁷

IKBS, however, had limitations, notably their inability to learn and their narrow focus. Manual updates were time-consuming. Currently, machine learning techniques have advanced to enable systems to learn autonomously. IKBS also lacked “common sense,” limiting their ability to handle unusual situations. This is a crucial point for India, where AI systems are being deployed in diverse and complex social contexts. For instance, a chatbot designed to provide government services needs to understand not only the specific query but also the underlying social and economic context of the user. Recognizing this, projects like CYC (Lenat & Guha, 1991) attempted to represent common sense knowledge by building an ontology of common- sense concepts. However, CYC encountered difficulties, particularly in handling the ambiguities of human language. More recent approaches leverage “big data” and open-source models to collect common sense knowledge from the web. These initiatives are highly relevant in India, where the availability of vast datasets and a growing digital population can be leveraged to train AI systems that possess a better understanding of the nuances of Indian society. Projects like Watson Healthcare, used by numerous healthcare organizations globally, represent the advanced state of AI in healthcare. In India, this technology can play a transformative role in improving access to quality healthcare, particularly in rural areas where specialist doctors are scarce. For example, AI-powered systems can analysis medical images and patient data to assist doctors in making faster and more accurate diagnoses, leading to better patient outcomes.⁸

The Roots of Intelligent Machines 

• “Defining Intelligent Systems9” According to, John McCarthy, who coined the term in 1956, defined it as the science and engineering of creating intelligent machines. It encompasses the development of computer systems capable of tasks that typically require human intelligence, such as visual perception, speech recognition, decision- making, and language translation. Google’s predictive search is a prime example. By analysis user data like browsing history, location, and age, Google uses these systems to anticipate search The dream of creating thinking machines is an ancient one, echoing through classical myths of artificial beings. Alan Turing’s 1950 paper, “Computing Machinery and Intelligence10,” introduced the “Imitation Game” (now known as the Turing Test), posing the question of whether machines could truly think.

Despite decades of effort, a machine that can convincingly pass the Turing Test remains elusive.

The years following Turing’s paper saw the first tentative steps toward realizing this dream. In 1951, Marvin Minsky and Dean Edmunds11 built a rudimentary neural network, a system of 3,000 vacuum tubes simulating just 40 neurons. The following year, Arthur Samuel developed a pioneering machine learning program for checkers, demonstrating early potential in game- playing artificial intelligence. The field of artificial intelligence formally emerged in 1956. At the Dartmouth Conference, John McCarthy coined the term “artificial intelligence,” defining it as the science and engineering of making intelligent machines. ¹ The following years saw the establishment of dedicated AI research labs, including the influential MIT AI Laboratory, founded in 1959.

• Milestones of Progress: The 1960s witnessed further advancements. Robots began to appear on factory assembly lines, with the first deployment at General Motors in Early chatbots like ELIZA, created in 1961, offered a glimpse into the potential of natural language processing. A landmark achievement came in 1997 when IBM’s Deep Blue defeated world chess champion Garry Kasparov. The 2005 DARPA Grand Challenge, won by the self-driving robotic car Stanley, showcased the progress in autonomous systems. In 2011, IBM’s Watson, a question-answering system, triumphed over human champions on the game show Jeopardy!
• The Age of Big Data and Intelligent Systems: Today, we live in the age of Big Data. The sheer volume of information available presents both a challenge and an opportunity. Intelligent systems have become essential tools across diverse sectors, from technology and finance to healthcare and entertainment. The availability of vast datasets, coupled with increasing computing power, allows even relatively simple algorithms to achieve impressive results through brute-force learning. While the pace of Moore’s Law may be slowing, the exponential growth of data continues to fuel progress in the field.

The future of artificial intelligence is full of promise. Natural language processing is poised for significant breakthroughs, as evidenced by the rapid evolution of automated conversational systems. Self-driving vehicles are likely to become a common sight on our roads in the coming decades. The long-term aspiration for some researchers is the development of artificial general intelligence (AGI)12—machines that possess human-level cognitive abilities across a wide range of domains. Whether AGI is achievable within the next 50 years remains an open question. Even if the technological hurdles are overcome, significant ethical considerations will need to be addressed before such systems are deployed. The development of advanced artificial intelligence raises complex societal and political questions that will require careful consideration. In the near term, the focus remains on enhancing the performance of intelligent systems, driven by the ever-increasing power of computation.

Conclusion and Suggestions

The legal landscape surrounding artificial intelligence is rapidly evolving, particularly in a diverse and dynamic nation like India. India’s approach to regulating AI is multifaceted, addressing critical areas like data protection, liability, and sector-specific applications. The Personal Data Protection Bill, 2023, represents a significant step, aiming to balance individual privacy rights with the data demands of AI development. This legislation, with its focus on data localization and consent, differs somewhat from the EU’s GDPR, highlighting India’s emphasis on data sovereignty. The complex issue of liability for AI-driven errors or harm is also under scrutiny. As AI systems become more autonomous, determining responsibility in cases of accidents or misdiagnoses becomes crucial. India is likely to draw upon international best practices as it navigates this challenging terrain. Furthermore, a sector-specific regulatory approach seems probable, allowing for tailored rules in areas like healthcare, finance, and transportation, recognizing the unique risks and opportunities within each domain. This mirrors trends in other jurisdictions, like the US, where sector-specific AI regulation is gaining traction. Intellectual property rights related to AI algorithms and software also require clarification to incentivize innovation and protect creators. The legal framework for ownership and licensing is still developing, and clear guidelines are essential for fostering a thriving AI ecosystem.

Comparing India’s approach with other global models offers valuable insights. The EU’s AI Act, with its comprehensive, risk-based regulatory framework, provides a strong contrast. The EU prioritizes fundamental rights and ethical considerations, placing stricter regulations on high-risk AI applications. The US, on the other hand, favors a more fragmented approach, with regulations emerging at both federal and state levels. This reflects a greater emphasis on promoting innovation, sometimes at the expense of uniform standards. China’s approach is distinct, characterized by substantial government investment in AI and a regulatory focus often driven by national security and social control objectives. These varied approaches highlight the global struggle to balance innovation with ethical concerns and societal needs. India’s approach, while still under development, reflects its unique context, seeking to harness the potential of AI while safeguarding individual rights and promoting responsible innovation. The evolving nature of AI technology necessitates adaptive legal frameworks, capable of addressing unforeseen challenges and ensuring that AI serves the broader interests of society. As AI continues to permeate various sectors, legal clarity and international cooperation will become increasingly vital for navigating this complex and transformative technological revolution.

Notes:

¹ MICROSOFT TECHNOLOGYU LICENSING, LLC v. THE ASSISTANT CONTROLLER OF PATENTS AND DESIGN (Delhi High Court, 2023)

² OPENTV INC v. The Controller of Patents and Designs (Delhi High Court, 2023)

³ McCarthy, J., Minsky, M., Rochester, N., & Shannon, C. E. (1955). A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence. AI Magazine, 26(2), 12-14.

⁴ Simon, H. A. (1996). The Sciences of the Artificial (3rd ed.). MIT press.

⁵ Darlington, K. (2000). The essence of expert systems. Prentice Hall.

⁶ Lenat, D. B., & Guha, R. V. (1991). Building large knowledge-based systems: Representation and inference in the Cyc project. Addison-Wesley.

⁷ Newell, A., & Simon, H. A. (1971). Human problem solving. Prentice-Hall.

⁸ Shortliffe, E. H., Buchanan, B. G., Feigenbaum, E. A., & Sridharan, N. S. (1975). Computer-based consultations in clinical therapeutics: rationale and implementation. Computers and Biomedical Research, 8(3), 267-299

⁹ McCarthy, J. (1956). Dartmouth Summer Research Project on Artificial Intelligence.

¹⁰ Turing, A. M. (1950). Computing machinery and intelligence. Mind, 59(236), 433-460.

¹¹ McCarthy, J., et al. (1955). A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence.

¹² Russell, S. J., & Norvig, P. (2021). Artificial Intelligence: A Modern Approach (4th ed.). Pearson.