A 5-point framework to analyze software patents

Key insights from Software Patents Review #1 can be distilled into a 5-point framework to enable a complete and thorough analysis of software patent applications. I reviewed 7 patents along these parameters, all of which are waiting to be examined. I aim to highlight the intent of the law, specifically Section 3(k) of the Indian Patents Act, and to visualize how each patent application stands under scrutiny in light of the intent of the law.

The five-point framework can be stated as 1. Does the claim pertain only to a software method, algorithm, or business method? 2. Is there a novel hardware element integral to the invention? 3. Does the claim align with Section 3(k)’s legislative intent? 4. Does the patent promote or hinder public interest and innovation (e.g., FOSS compatibility)? 5. Final Assessment

In light of this framework, let us look at 7 patent applications. These patents were applied for or granted within the period of 01 January 2024 and 01 January 2025. All data can be obtained by searching the IP Search portal on the Intellectual Property India website with the dates as mentioned above and “Artificial Intelligence” as a keyword of interest in the abstract input field.

1) Title - Artificial Intelligence-Based System and Method for Sustainable Urban Planning (Application Number - 202441024950) (Status - Published) This patent application claims to have developed a novel software technique, employing pre-existing technological paradigms and frameworks in the space of data analytics and predictive modeling in optimizing and improving sustainable urban planning. Employing the Framework devised above, let us take a deeper look at this patent application and its eligibility - 1. Does the claim pertain only to a software method, algorithm, or business method? * The claim pertains entirely to a novel software technique that leverages AI algorithms and machine learning techniques in the service of sustainable urban planning. * While it does make mention of relying on third-party data sources like sensors, and satellite data, at no point in the application, does it make mention of novel hardware components specifically designed and implemented for the purpose of this patent. It seems as if the data is being collated through a software interface that is then employed for the patent. 2. Is there a novel hardware element integral to the invention? * The invention makes use of existing hardware technologies (mentioned as sensors, satellite imagery, etc.) but introduces no hardware innovations. 3. Does the claim align with Section 3(k)’s legislative intent? * It appears that the claim does not fully align with Section 3 (k)’s intent. Since the section excludes software development except in conjunction with novel hardware elements, it will fall outside the ambit of patentability as intended by the law. 4. Does the patent promote or hinder public interest and innovation (e.g., FOSS compatibility)? * Based on the patent application, it appears that granting the patent application will have the potential to hinder innovation, especially in the domain of open-source software. It could potentially restrict the development of newer AI-based sustainable urban planning solutions. * A second objection on the grounds of public interest is that the granting of the patent and its availability being made by way of licensing would potentially limit its broader access, harming the public interest. 5. Final Assessment - * The patent application is entirely devoted to a software methodology, not a novel hardware application * It does not align with Section 3 (k)’s intent. * It does not seem to promote public interest and innovation in Open-source software.

2) Title - Intelligent Task Scheduling and Load Balancing in Cloud Computing Systems via Swarm Intelligence Techniques (Application Number - 202441101106) (Status - Published) This patent application appears to pertain to the application of specific artificial intelligence methods, namely swarm intelligence algorithms, to the domain of optimizing cloud computing systems. Employing the 4 Questions Framework devised above, let us take a deeper look at this patent application and its eligibility -

1. Does the claim pertain only to a software method, algorithm, or business method?
   • The claim appears to pertain entirely to mathematical techniques and software programs by means of employing swarm intelligence algorithms to optimize task scheduling and load balancing in cloud computing systems.
2. Is there a novel hardware element integral to the invention?
   • There appears to be no novel hardware component to this invention. It appears to leverage existing hardware infrastructure as exists in the space of cloud computing and is entirely a software innovation.
3. Does the claim align with Section 3(k)’s legislative intent?
   • It does not appear to align with the intent of Section 3(k) of the Indian Patents Act, which excludes software patents per se from patentability.
   • Since the applications appear to deal entirely with the application of existing AI techniques like swarm intelligence algorithms (PSO, ACO, ABS, etc.) to a specific domain of computer science, i.e., cloud computing, it risks being categorized as a software-based invention rather than a patentable technical invention. The algorithms themselves might be seen as abstract ideas rather than providing a technical contribution.
4. Does the patent promote or hinder public interest and innovation (e.g., FOSS compatibility)?
   • Granting a patent in this case has the potential to limit innovation in the cloud computing space, especially since its software-based approach could hinder further development in the open-source ecosystem. Smaller entities or startups may be completely blocked out of innovations in the cloud computing space since they rely on open-source solutions.
5. Final Assessment -
   • Since the claim pertains to a software-only method, it could be excluded under Section 3(k).
   • Granting a patent would limit further innovation in cloud computing in the open-source domain.

3) Title - ARTIFICIAL INTELLIGENCE BASED LOAN PREDICTION FOR BANKING SECTOR (Application Number - 202441100740) (Status - Published) This patent application claims to have applied a specific class of algorithm within the domain of machine learning to the problem of classification of loan applicants in the banking sector as high-risk or low-risk. Employing the 4 Questions Framework devised above, let us take a deeper look at this patent application and its eligibility - 1. Does the claim pertain only to a software method, algorithm, or business method? * Since the claim pertains solely to the application of a pre-existing specific class of machine learning algorithms (Support-Vector Machines (SVMs)) to the problem of loan-applicant risk classification, it does not have a non-obvious technical or hardware solution. 2. Is there a novel hardware element integral to the invention? * The application is purely software-oriented and does not have any novel hardware components. It does not contain any new physical devices or tangible technological improvements or any signifier of hardware innovation. 3. Does the claim align with Section 3(k)’s legislative intent? * It appears that it would not align with Section 3(k)’s legislative intent, according to which patents are not to be granted for mathematical ideas, business methods, or software. Since the patent application deals primarily with the application of a mathematical algorithm (SVM) to the problem of risk classification, mediated through software, it would directly conflict with Section 3(k)’s intent to exclude patents for purely software innovations. 4. Does the patent promote or hinder public interest and innovation (e.g., FOSS compatibility)? * The granting of this patent would severely hinder public interest and innovation in the open-source ecosystem since it would restrict the use of a widely used machine learning algorithm that benefits the financial and open-source ecosystems. 5. Final Assessment - * Since the claims deal entirely with software methods and mathematical algorithms without providing any contribution beyond what already exists in the domain of machine learning and without any conjunction with new hardware solutions, it does seem likely to pass the test of patent eligibility as elucidated by the law.

4) Title - SYSTEM AND METHOD FOR NATURAL LANGUAGE PROCESSING IN CLOUD-BASED ARTIFICIAL INTELLIGENCE ASSISTANTS (Application Number - 202441100200) (Status - Published) This patent application deals with the software methods pertaining to natural language processing within cloud-based AI assistants. Employing the 4 Questions Framework devised above, let us take a deeper look at this patent application and its eligibility -

1. Does the claim pertain only to a software method, algorithm, or business method?
   • This patent application deals solely with software methods for dealing with natural language processing for cloud-based AI assistants. It primarily focuses on the integration of NLP techniques rather than the creation of new hardware solutions.
2. Is there a novel hardware element integral to the invention?
   • There is no novel hardware element integral to the invention. The entire focus is on the creation of new software modules integrating different NLP techniques with cloud computing systems in order to generate responsive AI assistants. Although it describes a hybrid processing architecture, the claim is entirely software-centric and does not describe a single physical hardware element.</li>
3. Does the claim align with Section 3(k)’s legislative intent?
   • The patent application does not align with Section 3(k)’s legislative intent which excludes the granting of patents to mathematical methods, software, or business methods. Since the application deals solely with software systems with no mention of novel hardware elements, it should be viewed as a purely software method.
4. Does the patent promote or hinder public interest and innovation (e.g., FOSS compatibility)?
   • Since the patent application utilizes pre-existing NLP work and applies it to cloud-based AI assistants, granting a patent could cause significant hindrances to future innovation in this domain. This sort of work is foundational and granting a patent would restrict further developments for smaller entities and open-source ecosystems.
5. Final Assessment -
   • The claim lacks a novel hardware component and focuses on pure software/mathematical algorithms, making it likely to fall under Section 3(k), which excludes software-based inventions from being patentable in India.

5) Title - SYSTEM AND METHOD FOR ENTERPRISE DOCUMENT PROCESSING USING COGNITIVE INTELLIGENCE BASED FEATURE ENGINEERING\ (Application Number - 202441100579) (Status - Published) This patent application deals with software and mathematical methods and algorithms used in enterprise document processing. It employs a cognitive intelligence approach for structuring and analyzing the data and primarily uses software functions employing mathematical algorithms and linguistic intelligence for this task. Employing the 4 Questions Framework devised above, let us take a deeper look at this patent application and its eligibility -

1. Does the claim pertain only to a software method, algorithm, or business method?
   • The claims pertain entirely to software and mathematical methods for enterprise document processing. This method involv\es using a multi-layered cognitive intelligence approach to structuring and analyzing the data employing various software methods for this purpose.
2. Is there a novel hardware element integral to the invention?
   • This system does not involve any new hardware element, and the system focuses entirely on software-based modules and algorithms for document analysis, which can run on a server or cloud-based infrastructure.
3. Does the claim align with Section 3(k)’s legislative intent?
   • This patent application does not align with Section 3(k)’s intent since it appears to have been based on standard cognitive intelligence techniques (NLP, mathematical analysis, and domain-specific processing), with no hardware innovation.
4. Does the patent promote or hinder public interest and innovation (e.g., FOSS compatibility)?
   • If patented, this application would significantly hinder public interest and innovation by placing a limit on the development and use of common document processing techniques in the AI and cognitive computing domain. These methods are widely used in research, academia, and industry and a patent would place a hindrance over their usage.
5. Final Assessment -
   • Since the system does not include a novel hardware element or a technical advancement beyond existing methods in the domain of cognitive intelligence and document analysis, it likely does not meet the patentability requirements under Section 3(k).

6) Title - AN EEG-BASED SYSTEM AND METHOD FOR ALZHEIMER'S DIAGNOSIS USING MACHINE LEARNING AND EXPLAINABLE ARTIFICIAL INTELLIGENCE (Application Number - 202441100081) (Status - Published) This patent application deals with the application of software methods and algorithms to the diagnosis of Alzheimer’s disease. Employing the 4 Questions Framework devised above, let us take a deeper look at this patent application and its eligibility -

1. Does the claim pertain only to a software method, algorithm, or business method?
   • Yes, it appears to pertain only to a combination of software methods and algorithms, specifically employing signal processing, to the diagnosing of Alzheimer’s disease using EEG signals.
2. Is there a novel hardware element integral to the invention?
   • The patent application does not appear to describe any novel hardware element. Although it does explain the collection of EEG signals using standard technology in use today, its focus primarily lies in its employment of software-based processing of the EEG signals using machine learning algorithms in the field of signal processing. The hardware used is routine and does not signify any novel innovation.
3. Does the claim align with Section 3(k)’s legislative intent?
   • It does not appear to align with Section 3(k)’s legislative intent since this section excludes software, algorithms, and business methods unless there is a novel hardware component or a technical solution.
4. Does the patent promote or hinder public interest and innovation (e.g., FOSS compatibility)?
   • The patent application does not align with Section 3(k)’s legislative intent which excludes software components unless in conjunction with novel hardware innovation.
   • The usage of machine learning and signal processing in EEG-based diagnosis is fairly common in research in the healthcare industry. Granting a patent will prove deleterious by imposing a licensing regime over a subject that is extensively used by researchers, medical practitioners, and engineers in the healthcare sector.
5. Final Assessment -
   • The system and method for diagnosing Alzheimer's disease (AD) using EEG and machine learning techniques primarily describe a software-based algorithm that involves signal processing, machine learning classifiers, and explainable AI.
   • Therefore, the patent should not be granted as it is primarily a software method and algorithm without sufficient novelty or hardware innovation

7) Title - A SYSTEM FOR AUTONOMOUS WATER HYACINTH MANAGEMENT USING COMPUTER VISION, ARTIFICIAL INTELLIGENCE, AND IOT (Application Number - 202441024950) (Status - Granted) This patent application was granted and deals with an innovative system that blends a unique combination of hardware and software components to deal with autonomous water hyacinth management. Employing the 4 Questions Framework devised above, let us take a deeper look at this patent application and its eligibility -

1. Does the claim pertain only to a software method, algorithm, or business method?
   • The system comprises both software methods and innovative hardware components such as a boat body, detection unit (sensors, camera, microcontroller), movement unit (motor control), and a collection module (conveyor belt, crusher module).
   • It combines software for the control and tracking of tangible hardware components.
2. Is there a novel hardware element integral to the invention?
   • The invention includes specific hardware components such as
     • Turbidity and temperature sensors for water quality assessment.
     • A camera module for visual identification.
     • A conveyor belt system for collecting water hyacinth.
     • A crusher module for resizing collected material.
     • A microcontroller and central processing unit to integrate and coordinate hardware functions
3. Does the claim align with Section 3(k)’s legislative intent?
   • The claim aligns with Section 3(k)’s legislative intent since it does not focus solely on a “computer program per se” but rather integrates software with hardware to achieve a technical solution.
   • The invention goes beyond a standalone computer program by incorporating physical components to address a real-world problem.
4. Does the patent promote or hinder public interest and innovation (e.g., FOSS compatibility)?
   • The system promotes the public interest by providing a practical solution to the environmental issue of water hyacinth proliferation. By automating detection, tracking, and removal, the invention reduces labor costs and offers an efficient alternative to manual management.
5. Final Verdict
   • This invention appears to meet the patentability criteria:
     • It integrates novel hardware and software to address a technical challenge.
     • It does not fall under “computer programs per se” or other exclusions under Section 3(k).
     • The invention contributes to the public interest by tackling environmental problems in an efficient and autonomous manner