11 J Global Clinical Engineering Vol.7 Issue 2: 2025 Received July 31, 2024, accepted March 4, 2025, date of publication April 11 2025. Review Artificial Intelligence-Driven Insights for Regulatory Intelligence in Medical Devices: Evaluating EMA, FDA, and CDSCO Frameworks Radhakrishan Gaur1, Afifa Akram1, Indu Singh2 and Shikha Baghel Chauhan3* 1 Department of Drug Regulatory Affairs, Amity Institute of Pharmacy, Amity Institute of Pharmacy, Amity University, Noida, UP-201301, India. 2 Department of Pharmaceutics, Amity Institute of Pharmacy, Amity University, Noida, UP-201301, India. 3 Department of Industrial Pharmacy, Amity Institute of Pharmacy, Amity University, Noida, UP-201301, India. * Corresponding Author Email: schauhan@amity.edu ABSTRACT The current review elaborates Artificial Intelligence (AI) in medical devices is changing the landscape of diagnostics allowing for more accurate and efficacious treatments leading to better patient care. An overview of AI technologies and their applica- tion in medical devices elaborates on AI technologies, such as neural networks and advanced data analytics being applied in diagnostic imaging and patient-monitoring preventative analytic models. Machine learning, a subset of AI, enables devices to learn from data and improve their performance over time, enhancing diagnostic accuracy and personalized treatment plans. An elaborated critical review is presented for the regulatory strategies implemented by relevant global leaders, such as the European Union (EU), the United States (US Food and Drug Administration, FDA), and India (Central Drugs Standard Control Organization of India, CDSCO). This is indicative of the EU regulatory approach as observed through reflection paper by the European Medicines Agency (EMA) on a methodology to assess AI technologies used in conjunction with medicinal products, and the Software as a Medical Device (SaMD) guideline by the FDA in the United States. The discussion is on adaptive regulatory strategies, an overview of some pre-certification programs, and detailed advice to manufacturers about compliance with the processes. Also, India aligning with the International Medical Device Regulators Forum (IMDRF) guidelines shows its appetite to help build an extensive regulatory framework for AI-powered medical devices. The current review concludes by highlighting the need for continued coordination between regulators, manufacturers, and healthcare players so that AI advances are safe and adherent to the regulations that improve overall patient care. Keywords—Regulatory framework, Patient monitoring, Diagnostics, Neural networks, Machine learning. Copyright © 2025. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY): Creative Commons - Attribution 4.0 International - CC BY 4.0. The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduc- tion is permitted which does not comply with these terms. http://www.globalce.org http://globalce.org http://globalce.org mailto:schauhan@amity.edu https://creativecommons.org/licenses/by/4.0/ https://creativecommons.org/licenses/by/4.0/ Gaur, Akram, Singh, Chauhan: Artificial Intelligence-Driven Insights for Regulatory Intelligence in Medical Devices: Evaluating EMA, FDA, and CDSCO Frameworks J Global Clinical Engineering Vol.7 Issue 2: 2025 12 INTRODUCTION The medical device industry is being revolutionized by Artificial Intelligence (AI) through more efficient, data- driven, and adaptive regulatory practices. To improve pro- cesses, decision-making ability, and quality of healthcare provision, regulatory bodies around the world, such as the European Union (EU), US Food and Drug Administration (US FDA), and Central Drugs Standard Control Organization of India (CDSCO), are employing AI-based technologies. AI-driven medical devices extract valuable insights from medical data by using advanced methods of data analysis as well as machine learning (ML) algorithms. They then assist doctors in making better decisions, which have a positive impact on patients’ health outcomes.1 However, for medical devices guided by AI to be safe and efficient, they should adhere to Good Machine Learning Practices (GMLP), real-world performance tracking, compliance monitoring, and AI-ML software used as medical devices. Other essential components include regulatory frameworks, AI-powered regulatory documentation, AI-assisted regu- latory expertise, and AI-driven regulatory insights.2 The integration of these elements ensures that manufacturers of medical devices can safely and efficiently develop, test, and deploy AI-enabled devices.3 This approach improves patient outcomes and enhances healthcare quality, as it affects all aspects of health services, including expenses, as we observe in the text. The FDA is at the forefront of developing guidelines for evaluating the safety and ef- fectiveness of AI-enabled medical devices worldwide. The use of AI in healthcare raises a number of complex regulatory issues, some of which are unique to AI technol- ogy. A key challenge is the transparency of AI algorithms, which sets them apart from more traditional regulated technologies.4 To address these challenges, regulatory bodies are developing training programs and guidelines to enhance regulatory expertise in AI and ML. Overview of Artificial Intelligence in Medical Devices Artificial Intelligence means a system that acts with the help of machines and predicts or suggests actions in real-world but potentially also in virtual environments according to the objectives set by humans. These systems require both ML and human-based input to identify, translate, and create the context around, and perceive real or fictive environments. It parses these perceptions automatically, and then abstracts them into models. The data are then leveraged to create models for mining infor- mation, or possible decisions are modeled by the process of model inference.5 The applications of medical devices have been listed in Figure 1. Applications of AI in Medical Devices FIGURE 1. Application of AI in medical devices. Disease detection and diagnosis: AI algorithms, in- cluding deep learning models such as convolutional neural networks, can help doctors make decisions in many areas of medicine, such as oncology, radiology, ophthalmology, and general medicine. These algorithms analyze medical images (e.g., MRI, CT scans, X-rays, etc.) and other patient information to assist in diagnosis. Studies have shown that these models can reduce waiting period, enhance medication compliance, and tailor insulin doses, among other benefits. AI for personalized medicine: AI tools can recom- mend the most effective treatments based on genetic and medical history, as well as lifestyle factors, leading to more precise and effective treatments. Continuous learning and improvement: AI devices can learn from real-world use to improve their perfor- mance over time, adapting to new clinical scenarios and enhancing their detection capabilities, which in turn improve patient care. 13 J Global Clinical Engineering Vol.7 Issue 2: 2025 Gaur, Akram, Singh, Chauhan: Artificial Intelligence-Driven Insights for Regulatory Intelligence in Medical Devices: Evaluating EMA, FDA, and CDSCO Frameworks Enhanced diagnostics: AI-powered diagnostics involve analyzing multiple medical data sources to improve early and accurate diagnosis of illnesses. This includes mining electronic health records (EHRs), automated laboratory report analysis, and medical image interpretation (e.g., radiology, pathology, etc.). Performance evaluation: AI devices can be helpful in the performance evaluation process by increasing efficiency, improving patient outcomes, and simplifying data collection. This includes the use of AI in post-market surveillance and real-world performance monitoring to ensure the ongoing safety and efficacy of medical devices.6 Predictive analytics for monitoring: AI can be used in wearable devices to monitor patient vitals and provide predictive insights on conditions such as heart failure and glucose levels. Regulatory compliance and clinical trials: AI can au- tomate data generation, validation, and metadata analysis for regulatory submissions, enhancing clinical trials by predicting outcomes, optimizing patient recruitment, and ensuring compliance with global regulatory standards. AI for diagnostics: ML models in diagnostic devices can aid in the automatic interpretation of test results, such as ECG and blood tests. Natural Language Processing (NLP) for documenta- tion: AI can automate and streamline clinical documenta- tion, making it easier to comply with medical regulations and ensure accurate patient records. Robotic surgery: ML algorithms can assist surgeons with precision during operations. AI in Software as a Medical Device (SaMD): AI is being integrated increasingly into SaMD to enhance functionality and performance. Importance of Regulatory Intelligence and Compliance Monitoring Regulatory intelligence is essential for compliance monitoring, helping companies to keep updated about changing regulations. It ensures that the company remains in accordance with all laws and requirements. Regulatory intelligence empowers businesses in anticipating and deal- ing with global changes in regulations, without having to monitor manually different sources of regulatory data. A proactive approach ensures organizations respond faster to make the changes they need, mitigating risk and reducing costs. For example, regulatory intelligence platforms have tools for tracking changes in the rules and analyzing their impact to discover which obligations might stem from a change of laws that may apply specifically depending on what is being done by you.7 These tools ensure the regu- latory updates and reports generated on the platform itself, making it more efficient for operations at lower cost. For instance, the use of AI in regulatory intelligence is expected to automate monitoring and dissemination in the future, freeing up regulatory experts to concentrate on high-value tasks and decision-making.8 Regulatory information tools for monitoring consist of regulatory agencies, website monitoring tools, news organizations, and subscription-based services regulatory intelligence for regulatory strategies and operations, product due dili- gence (pre- or post-deal), target products identification, and clinical development to regulatory submissions. This capability helps organizations stay current with applicable regulations, understand the implications, and create an intelligence report that can be used to remain compliant while reducing resource overhead.9 Need for Effective Regulatory Frameworks to Keep Pace with AI Innovations for Medical Devices The current regulatory framework for AI-enabled medical devices is characterized by a wide void of strict laws and guidelines. The FDA (United States), the Medicines and Healthcare products Regulatory Agency (MHRA, UK), and the Health Canada have published preliminary thoughts; however, regulations by law are only enforceable when issued. Only Saudi Arabia has implemented proscriptive guidance so far. This ambiguity obliges manufacturers to employ regulations not modernized for the new and revised short-cycle drugs, serving as roadblocks toward algorithmic adaptability along with dynamic data-driven updates by treating newly acquired information without constraints.10 Gaur, Akram, Singh, Chauhan: Artificial Intelligence-Driven Insights for Regulatory Intelligence in Medical Devices: Evaluating EMA, FDA, and CDSCO Frameworks J Global Clinical Engineering Vol.7 Issue 2: 2025 14 essential when utilizing CE-certified devices in a clinical trial to safeguard the rights, safety, and well-being of participants as well as the integrity, and applicability of evaluation data. It is clear from the reflection paper that EMA scrutinizes whether the characteristics of medical devices are suitable for generating evidence needed for marketing authorization application, or whether a device provides recommendations in the Summary of Product Characteristics (SmPC).15 EMA’s Approach to Evaluating AI Technologies in Medical Device Development and Authorization The EMA’s strategy for assessing AI Technologies in the development of medical products is as follows: The EMA has released a reflection paper that focuses on promoting the utilization of AI throughout the lifecycle of medicinal products. This covers medical devices that are used within clinical trials to provide evidence for a marketing authorization application or in case they are used with a medicinal product.16 The EMA assesses such devices to decide whether they can provide adequate evidence for approval in EU countries. These recom- mendations include information about how to conduct AI research, which needs to be updated regularly, given the advances in the field and the new knowledge generated by research. If SmPC recommendations, such as posology or monitoring incorporate advice from an AI-enabled medical device, all relevant aspects of that combination are considered during assessment by the EMA.17 According to this reflection paper, general guidelines and expecta- tions applicable for medical devices would also apply to the clinical trial and marketing authorization contexts using AI/ML-based approaches.18 Overall, the EMA is adopting a risk-based strategy, in- structing the sponsors to consider whether the AI system presents risks to patients, and if so, then to seek early regulatory advice from the EMA. The EMA is getting ready to examine applications that incorporate AI/ML systems into the lifecycle of medical products.19 Overview of the European Medicines Agency’s (EMA) Reflection Paper The EMA has released a preliminary document outlin- ing its current stance on the use of AI and ML to enhance REGULATORY APPROACHES TO AI IN MEDICAL DEVICES Artificial intelligence in medical device regulations is now evolving quickly, as the EU, the United States, and India are adopting new pathways to ensure that these technologies are safe and effective. The EU is developing a comprehensive AI Act, which sets out strict rules on any use of the technology, while the United States continues to rely on its existing regulations and guidelines. India is aligning its regulatory framework with international standards, particularly those set by the International Medical Device Regulators Forum (IMDRF).11 These regu- latory mechanisms play an important role in preventing the operational and planning data management system from becoming points of failure, and thus protect patients as well as healthcare practitioners by determining that medical devices powered with AI underperform reliably.11 European Union The EU wants to regulate AI in all areas, including healthcare, based on how dangerous it is. The proposed AI Act wants to make a legal definition of “AI system” and sets rules for how AI can be built into and used in medical devices. The EU AI Act makes sure that AI systems follow basic rights, safety rules, and morals by setting clear rules for how they can work.12 Overview of the European Medicines Agency’s (EMA) Reflection Paper The EMA has published a preliminary document discuss- ing the use of AI and ML in the entire lifespan of medicinal products, including those for human and veterinary use.13 The use of AI/ML systems for the clinical management of individual patients may result in the Medical Device Regulation (MDR) classifying in vitro diagnostics (IVDs) used in performance evaluation as medical devices, as stated in the reflection paper. A document issued by the Medical Device Coordination Group (MDCG) offers de- tailed guidelines on the qualification and classification of software as medical devices in accordance with the MDR, the in vitro Diagnostic Devices Regulation (IVDR), or both.14 Nonetheless EMA has no responsibility to clas- sify software by the rules. Additional prerequisites are 15 J Global Clinical Engineering Vol.7 Issue 2: 2025 Gaur, Akram, Singh, Chauhan: Artificial Intelligence-Driven Insights for Regulatory Intelligence in Medical Devices: Evaluating EMA, FDA, and CDSCO Frameworks the safe and efficient development, regulation, and use of human and veterinary medicines throughout the lifespan of the product. The EMA acknowledges the potential of AI to improve different areas of the pharmaceutical in- dustry, such as drug discovery, preclinical development, clinical trials, precision medicine, product information, manufacturing, and post-approval pharmacovigilance.20 Nevertheless, the agency underscores the importance of adopting a human-centered approach in all aspects of AI and ML development and implementation. It is crucial to adhere to the existing legal obligations, prioritize ethical considerations, and uphold fundamental rights. The EMA promotes transparency and comprehensibil- ity in the creation and verification of AI systems.21 This entails providing explicit documentation of the utilized data, applied algorithms, and achieved performance, with the level of explanation aligning with the level of risk. Sponsors should utilize reliable and accurate data when creating and testing AI systems, carefully choose and validate algorithms for specific purposes, establish continuous monitoring and maintenance plans to iden- tify and address any decline in performance over time, conduct thorough risk assessments, and take necessary measures to mitigate risks, and proactively collaborate with regulators to ensure compliance with AI-usage guidelines. The EMA’s preliminary reflection paper was available for public consultation until December 31, 2023.22 US Food and Drug Administration The US FDA regulates the utilization of AI in medical devices. The FDA evaluates AI/ML-enabled medical de- vices based on their use, using appropriate premarket pathways, such as 510(k) clearance and De Novo clas- sification for noncontroversial new technologies, and traditional PMA for innovative devices.23 The regulations governing medical devices are highly specific. The agency provides detailed recommendations and action plans for regulatory consideration in response to the challenges posed by the rapidly evolving field of AI technologies. These encompass the AI and Software as a Medical Device Action Plan, along with supplementary guidance on ML best practices, pre-specification change control plans, and transparency.24 FDA Guidelines for AI/ML-Based Software as a Medical Device Software as a medical device is a term coined by the IMDRF to refer to software that is specifically created for medical usage and can function on its own, without incorporating into a physical medical device.25 Overview of the FDA guidelines for AI/ML-based SaMD: The FDA is in the process of developing regulatory pathways for SaMD driven by AI and ML. These pathways include traditional premarket pathways, such as 510(k) clearance, De Novo classification, and Premarket Approval (PMA). Recognizing that AI/ML technologies are adaptive by design, the FDA has acknowledged that traditional regulatory approaches may not be sufficient for these devices. This is particularly important because the FDA no longer believes that the old requirement of locking algorithms post-training is adequate. Instead, they are moving toward a more adaptive framework that allows algorithms to undergo modifications under predefined change control plans, ensuring safety and effectiveness. This approach addresses the specific challenges posed by AI/ML medical devices.26 Important FDA Guidance Documents The document titled “Proposed Regulatory Framework for Modifications to AI/ML-Based Software as a Medical Device” published in April 2019 explores a potential method for evaluating AI/ML modifications before they are released to the market. In January 2021, the FDA issued the “AI/ML SaMD Action Plan”, which offers a comprehensive framework for regulating AI and ML technologies in SaMD.27 The title of the event in October 2021 was “Good Machine Learning Practice for Medical Device Development: Guiding Principles”. In April 2023, draft guidance was released on “Marketing Submission Recommendations for a Predetermined Change Control Plan for AI/ML-Enabled Device Software Functions.”28 The title of the publication in October 2023 was “Guiding Principles for Predetermined Change Control Plans in Machine Learning-Enabled Medical Devices.” The title of the publication released in June 2024 was Gaur, Akram, Singh, Chauhan: Artificial Intelligence-Driven Insights for Regulatory Intelligence in Medical Devices: Evaluating EMA, FDA, and CDSCO Frameworks J Global Clinical Engineering Vol.7 Issue 2: 2025 16 “Transparency for Machine Learning-Enabled Medical Devices: Guiding Principles.” 29 FDA Guidelines for AI/ML-Based Software as a Medical Device The FDA’s Total Product Lifecycle (TPLC) approach to regulation aims to expedite the enhancement process of SaMD utilizing AI and ML, while ensuring compliance with all essential safety protocols. Developers are anticipated to be transparent about the functionality of their products in real-world scenarios and adhere to both good ML practices (GMLP) and quality systems. The TPLC framework enables manufacturers to present change control plans that have been implemented prior to the initial premarket review process.30 This allows for specific modifications to be implemented without necessitating a fresh evaluation process. The FDA granted approval up to 2024 to approximately 64 medical devices that utilize AI and ML, with the majority being approved through the 510(k) pathway. The agency is currently endeavoring to enhance its regulatory framework in order to match the rapid pace at which these technologies are evolving.31 The recommendations from the guidelines: The FDA’s guidelines on AI/ML-based SaMD are classified into three broad categories, i.e., assuring the safety and effectiveness of these devices, promoting good ML practices in device development, and ensuring transparency about how algorithms reach their outputs.32 Most important suggestions are as follows: Thorough validation: AI and ML algorithms need to be thoroughly validated to show that they work as intended and give correct results. This would include clinical trials and tests against well-known ways of diagnosing. Risk management: Companies that create AI/ML algorithms are to search for and mitigate any risks associated with it. Biases in the data used to train an algorithm, errors made by the algorithm itself, and security vulnerabilities are all potential problems.24 Real-world performance monitoring: AI and ML models can learn from experience to improve performance over time, so the FDA recommends using real-world clinical data to continually assess a device’s safety and effectiveness. Good machine learning practices: According to the FDA, GMLP refers to following the ML development best practices throughout the entire ML lifecycle. If you follow a best practices process, such as training with great real-world data, keeping good model development and validation practices, or making sure the software has strong versioning support.33 Predetermined change control plans (PCCP): The AI/ML model should learn and change over time, so manufacturers should set up a PCCP that explains how it will do this. The PCCP should have steps for finding, evaluating, and lowering the risks that might come with changing algorithms.34 Clear user manuals: The SaMD should come with clear and detailed user manuals that explain what the device can and can't do and how the AI/ML algorithms work. Data transparency: Users should know what kind of data were used to train AI/ML models and if there are any biases that could affect the outputs of the device. Transparency of algorithmic functioning: The level of details about how the AI/ML algorithms work on the inside may change depending on how complicated is the device.35 FDA’s Regulatory Approach for Adaptive AI-Driven Devices While the FDA recognized that AI and ML may now be transformative in medical devices, it also recognizes that its traditional regulatory framework is likely not well suited for these ever-evolving techs. In response, the agency launched a new TPLC initiative that uses existing premarket pathways in combination with risk management and authorization from its Pre-Cert program. Nationwide strategies mean to diminish the obstructions confronting programming designers and other human service suppliers in executing conduct of science criticism; simultaneously, direction laid down with respect to a “strict control” change for guideline focus on ML techniques.36 With a large guide that provides transparency during each step of AI algorithms suggested by the FDA, this can help you make the most out of developing tools or models with features-proposed higher-up code. Two components of the TPLC framework, however, probably would require more statutory authority to fully implement than by EPA.37 India’s Central Drugs Standard Control Organization 17 J Global Clinical Engineering Vol.7 Issue 2: 2025 Gaur, Akram, Singh, Chauhan: Artificial Intelligence-Driven Insights for Regulatory Intelligence in Medical Devices: Evaluating EMA, FDA, and CDSCO Frameworks India’s Central Drugs Standard Control Organization CDSCO’s Alignment with IMDRF Guidelines The CDSCO is aligning its medical device rules with IMDRF guidelines. The CDSCO issued a notification on August 1, 2019, revising the Medical Devices Rules of 2017 to make SaMD subject to its existing regulatory framework. Following the IMDRF guidelines provides a road map to manufacturers in India by which they can comply with their regulatory obligations for SaMD prod- ucts. SaMDs are classified into risk-based classes (A–D), which are systematically defined as per the IMDRF risk framework.38 It determines the required level of regula- tory oversight and compliance. The CDSCO is working on creating a Digital Drugs Regulatory System (DDRS), which, however, is an AI-enabled open-source technol- ogy. The goal is to create a harmonious digital regula- tory environment based on international standards and best practices. As long as AI medical devices safeguard patient data and offer safe access to it for unauthorized parties as defined by new legislation, they can continue to be relevant in accordance with the ethical and open principles outlined by the IMDRF. The CDSCO has leaned into the IMDRF guidance as a way of developing its medi- cal device regulations, with an emphasis on technologies such as AI-powered SaMDs. The aim is to have a robust, risk-based, and globally harmonized regulatory system for medical devices in India.39,40 Overview of IMDRF guidelines for medical devices: The IMDRF has published a framework, which is adopted and applied by the CDSCO. Main safety and performance IMDRF regulations are as follows: IMDRF has released the document “Essential Principles of Safety and Performance for Medical Devices & IVDs” (IMDRF/GRRP WG/N47). This guidance provides high- altitude rules on designing and constructing medical devices to make them safe, by following which you can guarantee the safety characteristics that are in place when your device is used as intended. IMDRF/GRRP WG/N71:2021 Medical device regulatory review reports guidance regarding information to be included (IMDRF/GRRP, 2020). The guidance is to provide common format requirements for the content and format of a regulatory submission dossier for a medical device.41 This report encompasses crucial sections, such as the following: Regional administrative information Submission context Nonclinical evidence Clinical evidence Labeling and promotional material Quality management system information There are rules and principles for medical devices in India that were taken from the IMDRF and added by the CDSCO to the Medical Devices Rules, 2017. The CDSCO also publishes lists of medical devices that are categorized by risk. These lists match the IMDRF framework of classes A–D based on risk.42 Implications of CDSCO’s alignment with IMDRF guide- lines for AI-based medical devices in India: In India, the CDSCO has mostly made its rules about medical devices, such as SaMDs, the same as the rules and guidelines set by the IMDRF. The CDSCO has implemented the IMDRF’s risk-based system to categorize SaMDs into four groups based on their level of risk: low-risk (Class A), low–moderate risk (Class B), moderate risk (Class C), and high-risk (Class D). Essential principles of safety and performance: The CDSCO makes sure that SaMDs follow the IMDRF’s “Es- sential Principles of Safety and Performance of Medical Devices and IVD Medical Devices” when they design and make their products. Labeling and advertising: In India, SaMDs must follow the IMDRF’s “Principles of Labeling for Medical Devices and IVD Medical Devices” to make sure users get the right information. Structure of regulatory submissions: The CDSCO has used the IMDRF’s “Medical Device Regulatory Review Report” template to decide the information that should be included and how it should be organized in regulatory submissions for SaMDs. Gaur, Akram, Singh, Chauhan: Artificial Intelligence-Driven Insights for Regulatory Intelligence in Medical Devices: Evaluating EMA, FDA, and CDSCO Frameworks J Global Clinical Engineering Vol.7 Issue 2: 2025 18 TABLE 1. Comparative analysis for Regulatory framework for AI across various Regulatory agencies. Feature Sub-feature European Union (EU) United States (FDA) India (CDSCO) Focus Regulatory framework for AI Draft reflection paper on AI in medicinal products lifecycle defines “AI system” and proposes regulatory framework. EU AI Act focuses on general AI development and ethical considerations. AI/ML action plan outlines focus on developing guidance for AI/ML in medical devices.43 Aligns with IMDRF principles for AI-based medical devices.44 Traditional premarket pathways Existing CE marking framework with risk-based assessment considering AI components. Existing 510 (k) clearance, de novo classification, or premarket approval (PMA) pathways for SaMD. Follows risk-based classification similar to the US FDA. Post-market surveillance Requires manufacturers to have a post- market surveillance plan, potentially incorporating AI for anomaly detection and trend analysis.45 Requires manufacturers to monitor and report adverse events for SaMD, potential use of AI for real-world performance monitoring. Requires manufacturers to submit periodic safety update reports (PSURs), potential for AI-assisted data analysis. AI-specific requirements Transparency and explainability Emphasis on transparency in AI decision-making processes and explainability of results. Focus on good ML practices (GMLP) for development and validation of AI models. Aligning with IMDRF principles for data access, security, and responsible use.45 Change management The draft guidance highlights the necessity for predefined change control plans for AI systems to maintain continuous regulatory compliance. Guidance on predefined change control plans (PCCP) for AI/ML outline processes for managing changes in AI models. Aligns with IMDRF principles for managing changes in AI-based medical devices. AI advantages for regulatory bodies Efficiency and monitoring Potential for AI to improve efficiency in monitoring medical device performance and identifying potential risks. Potential for AI to streamline data analysis and automate risk assessment processes for SaMD. Potential for AI- assisted automation in regulatory processes once frameworks are established. Challenges Regulatory uncertainty Lack of clear and finalized regulations for AI in medical devices creates uncertainty for developers and regulators.46 Adapting existing frameworks to address the continuous learning nature of AI models presents challenges.47 Limited experience and resources for implementing AI-based regulatory processes.48 19 J Global Clinical Engineering Vol.7 Issue 2: 2025 Gaur, Akram, Singh, Chauhan: Artificial Intelligence-Driven Insights for Regulatory Intelligence in Medical Devices: Evaluating EMA, FDA, and CDSCO Frameworks Standards that are the same everywhere: In India, SaMDs must follow international rules, such as IEC 62304 for software lifecycle, IEC 60601-1 for embedded software, and IEC 82304-1 for standalone SaMDs.42 Table 1 provides a comparative analysis of the regulatory frameworks for AI across various regulatory agencies, highlighting key similarities and differences in their approaches. POSSIBLE INCLUSION AND ADVANCEMENT Artificial intelligence and innovative regulatory intel- ligence integration with medical devices are to revolution- ize efficiency, accuracy, and compliance. The changing landscape of the medical device regulatory framework is driven by the growth in AI solutions, which are now utilized by prominent regulators, such as the EMA, FDA, and CDSCO. This section discusses potential progress and distinct types of innovation that are tailored to each regulator’s specific issues. Focus areas include better data integration and interoperability, regulator-specific AI applications via innovation, and an ethical and trans- parent framework for AI development. The purpose is to present AI’s critical and ever-growing role in achieving better regulation, patient safety, and true international alignment. Enhanced Data Integration and Interoperability When it comes to regulating medical devices, it’s very important to combine data from different sources, such as clinical trials, post-market surveillance, and EHRs. This is because these sources provide important information for making sure that all regulations are followed. In response to this need, AI algorithms are created that allow huge amounts of unstructured data from many systems to be collected, processed, and turned into useful information. These algorithms can extract and align large amounts of data in real time, while most other ways of integrating data are done manually.49 One example of this trend is the use of ML models to connect patterns and changes in different databases. This allows regulators to see how the device works in a bigger picture for safety reasons. Also, it’s faster to share data with regulatory bodies and other important parties, thanks to the progress in standardizing and connecting data. Health Level Seven International (HL7) standards or Fast Healthcare Interoperability Resources (FHIR) are examples of universal data standards and protocols that make sure that different systems talk to each other correctly so that data are used. Not only do these improvements speed up the regulatory process, but they also make it easier to spot and act on safety signals quickly. Data integration and interoperability made pos- sible by AI help regulatory agencies make smarter rules to protect patients and keep up with MedTech innovation.49 Regulatory Body-Specific AI Innovations Customization of AI solutions to tailor specific regula- tory challenges encountered by EMA, FDA, and CDSCO is critical in streamlining regulatory operations and maintaining compliance acts. Every regulatory body works within different laws and healthcare contexts, call- ing for highly differentiated AI use cases that answer to individualized requirements. The EMA, for example, has been using AI to facilitate the review of clinical trial data processing validations with a view to speeding up the approval process swiftly and accurately while maintain- ing significant safety assessments.49 Through real-time post-market surveillance, the FDA has led the way in its application of ML algorithms and other AI technologies to track adverse event data (and even device performance) continuously. In India, the CDSCO is turning to AI-powered ML tools that can help improve regulatory submission and approval time with so many medical devices entering a diverse market rapidly.49 These customized AI solutions have seen a level of advancements proven by successful implementation by regulatory bodies. The FDA’s Sentinel Initiative alone is a case of how large-scale healthcare data are analyzed for post-market safety surveillance, bringing the process to detect potential safety signals by significantly reduc- ing time with AI. This is the use of AI in implementing adaptive pathways by EMA, leading to quicker entry for innovative medicines into the hands (health) of patients while upholding high safety standards. For instance, in India, AI-driven platforms are developed to automate the CDSCO regulatory review process to save substantial time and effort on administrative tasks.49 Thus, these customized AI solutions are able to solve specific regulatory problems and also lay the foundation Gaur, Akram, Singh, Chauhan: Artificial Intelligence-Driven Insights for Regulatory Intelligence in Medical Devices: Evaluating EMA, FDA, and CDSCO Frameworks J Global Clinical Engineering Vol.7 Issue 2: 2025 20 for agility in regulation. Thus, this review article exhibits these progressions and contextual analyses which are used to understand the implementation of AI by regulatory bodies that augment their processes based on efficiency criteria for patient safety. This observation illustrates just how disruptive AI could be in regulatory intelligence, and why constant innovation is essential to adapt its capabili- ties to the evolving demands of their regulations.49 The FDA’s Pre-Cert Program is designed to enhance regulatory efficiency and innovation. Pre-certified compa- nies can release updates and new products more quickly, allowing for faster adoption of innovative technologies. This program also provides flexibility in evolving with technology, particularly for AI/ML, while maintaining a focus on safety through post-market monitoring and company excellence reviews.49 Ethical and Transparent AI Progress in the evolution of responsible AI frameworks has become more urgent because medical devices are the area where advanced regulatory decisions are made, at least partially by an AI system. This is an additional measure important to ensure transparency and accountability, in turn ensuring the trustworthiness of AI-driven processes for public health maintenance. Ethical AI frameworks are created to enable the development of these systems, so that they are designed and implemented in line with ethical principles as well as regulatory standards.49 This consists of having protocols for data privacy, bias mitiga- tion, and ensuring equitable access to All-driven insights. These frameworks also underscore the need for account- ability—that people must know who is responsible in cases where AI systems produce harm. The use of explainable AI (XAI) models provides one of the most important advancements in this space. In cases where traditional AI systems are essentially “black boxes,” withholding any information about their decision-making process, XAI models aim to better unpack the results of our AI systems. Regulators and stakeholders should be able to unpack the decision criteria upon which the AI has based its conclusions, so that decisions at any time can still accompany handling scrutiny. For example, sub- jecting an XAI model to a regulatory body looking at the integrity of any selected medical device should be able to understand how exactly it arrived at its decision based on concrete data and patterns. The FDA is also showing initiatives in improving more ethically aligned and transparent AI through projects such as those exploring ways to bring XAI into its regula- tory review pathways. By requiring that AI models used in regulatory submissions be interpretable, the FDA is ensuring that those tools can have both regulator and patient trust. On the other hand, some form of guidance and transparency in strategy, be it through routine audit processes for addition or on-account validation built-in algorithms, at EMA and CDSCO are likely to hold posi- tion.49 The FDA is also showing initiatives in improving more ethically aligned and transparent AI through proj- ects such as those exploring ways to bring XAI into its regulatory review pathways. By requiring that AI models used in regulatory submissions be interpretable, the FDA is ensuring that those tools can have both regulator and patient trust. On the other hand, some form of guidance and transparency in strategy, be it through routine audit processes for addition or on-account validation built-in algorithms, at EMA and CDSCO are likely to hold position.49 One World, One Regulation In an era where AI technologies are rapidly advancing and transforming the landscape of medical devices, the need for a harmonized global regulatory framework has become increasingly evident. The vision of “One World, One Regulation” encapsulates the aspiration for a unified approach to ensure that AI-powered medical devices meet consistent standards of safety, effectiveness, and ethical use across all regions. This vision is not just about standardization but about fostering an environment where innovation can thrive without being hindered by disparate regulatory requirements. By aligning regula- tory approaches, we can facilitate the rapid deployment of AI technologies, making advanced healthcare solutions more accessible globally. Moreover, a unified framework builds trust in these technologies by ensuring rigorous oversight and consistent performance standards. It also promotes collaboration among regulatory bodies, stream- lining the approval process and reducing the burden on manufacturers. Ultimately, the “One World, One Regula- tion” approach is essential for realizing the full potential 21 J Global Clinical Engineering Vol.7 Issue 2: 2025 Gaur, Akram, Singh, Chauhan: Artificial Intelligence-Driven Insights for Regulatory Intelligence in Medical Devices: Evaluating EMA, FDA, and CDSCO Frameworks of AI in healthcare, ensuring that these technologies can be safely and effectively utilized to improve patient out- comes globally.49 CONCLUSION The regulatory framework for medical devices incorpo- rating AI is currently experiencing substantial changes. The EMA, FDA, and CDSCO are actively formulating frameworks and guidelines to guarantee the secure and efficient utilization of AI. These initiatives seek to tackle problems stemming from unclear regulations and a lack of expertise in the field. The EMA’s reflection paper emphasizes the importance of conducting a thorough risk analysis at the time of making decisions about the implementation of AI in medicines. This analysis should take into account factors such as the transparency and interpretability of the AI system as well as its ability to manage changes throughout its lifecycle. Artificial intelligence has the capacity to completely transform the field of drug discovery and regulatory processes, encompassing everything from the initial development of pharmaceuticals to ongoing monitoring after approval. The EMA undertakes a comprehensive as- sessment of these assertions. Furthermore, the scientific validity and applicability of medical devices incorporat- ing AI that are utilized in clinical trials within the EU are evaluated to ascertain their effectiveness and suitability in all member states. The EMA recommends that spon- sors assess the potential hazards of implementing new AI systems on patients and promptly seek guidance from the regulatory body. Additional suggestions involve the creation and evaluation of AI systems using clear and well-documented approaches, offering strong justifica- tions for AI implementation in particular situations, and comprehending the accompanying hazards and constraints. The FDA has released guidelines for SaMD that employ AI and ML. These guidelines provide recommendations for effectively utilizing ML, incorporating pre-established change control systems, and guaranteeing the openness of data and algorithms. The FDA acknowledges the revo- lutionary potential of AI and ML in medical devices while recognizing the distinct challenges posed by them. The adoption of the TPLC approach facilitates the ongoing enhancement and progression of AI/ML-based software as medical devices, ensuring the preservation of safety and efficacy throughout all premarket review activities. The CDSCO has implemented the regulations established by the IMDRF, encompassing the risk-based classification system and fundamental principles for guaranteeing safety and performance. The CDSCO follows IMDRF guidelines to ensure the performance and safety of medical devices and IVDs while designing and restructuring SaMD. Indian SaMD must adhere to global standards, including IEC 62304, for managing risks in the software life cycle, IEC 60601-1 for embedded software, and IEC 82304-1 for standalone SaMD. Although regulatory ambiguity and limited expertise pose significant challenges, these regulatory bodies are making efforts to ensure the safe and effective utilization of AI in medical devices, ultimately improving patient outcomes. The main goal of the EU AI Act is to effectively govern scientific progress and the extensive implemen- tation of AI in medical devices and other products and services. These organizations work diligently to ensure that AI systems comply with safety regulations and ethical considerations while also protecting fundamental rights. AUTHOR CONTRIBUTIONS Conceptualization, R.G.; Methodology, S.B.C.; Software, I.S.; Validation, I.S.; Investigation, S.B.C.; Resources, I.S.; Data Curation, A.A.; Writing–Original Draft Preparation, R.G. and A.A.; Writing–Review & Editing, S.B.C.; Visualiza- tion, S.B.C.; Supervision, I.S.; Project Administration, I.S. ACKNOWLEDGMENTS The technical assistance needed to finish this project was provided by Amity Institute of Pharmacy at Amity University Noida, for which the authors are grateful. FUNDING This research received no external funding. DATA AVAILABILITY STATEMENT Not applicable. CONFLICTS OF INTEREST The authors declare they have no competing interests. Gaur, Akram, Singh, Chauhan: Artificial Intelligence-Driven Insights for Regulatory Intelligence in Medical Devices: Evaluating EMA, FDA, and CDSCO Frameworks J Global Clinical Engineering Vol.7 Issue 2: 2025 22 ETHICS APPROVAL AND CONSENT TO PARTICIPATE Not applicable. CONSENT FOR PUBLICATION Not applicable. FURTHER DISCLOSURE Not applicable. REFERENCES 1. AI: Upcoming regulations for pharma and medical devices. Available online: https://www.scilife.io/blog/ artificial-intelligence-regulations. 2. Food and Drug Administration (FDA). Artificial intelligence and machine learning in software as a medical device. Available online: https://www.fda. gov/medical-devices/software-medical-device-samd/ artificial-intelligence-and-machine-learning-software- medical-device. 3. Shick, A.A., Webber, C.M., Kiarashi, N., et al. Transpar- ency of artificial intelligence/machine learning-enabled medical devices. NPJ Digit Med. 2024;7(1):21. https:// doi.org/10.1038/s41746-023-00992-8. 4. Parenteral Drug Association (PDA). A comprehensive review of regulatory intelligence: Exploring tools and program maturities. Available online: https:// www.pda.org/pda-letter-portal/home/full-article/a- comprehensive-review-of-regulatory-intelligence- exploring-tools-and-program-maturities. 5. Koli, P., Raut, S., Mande, R., et al. An overview of regulatory intelligence. Int J Creat Res Thoughts. 2023;11(4):b953–b964. http://ijcrt.org/viewfull. php?&p_id=IJCRT2304232. 6. Zhou, K. and Gattinger, G. The evolving regulatory paradigm of AI in Med Tech: A review of perspec- tives and where we are today. Ther Innov Regul Sci. 2024;58(3):456–464. https://doi.org/10.1007/ s43441-024-00628-3. 7. Wenzel, M. and Wiegand, T. Toward global valida- tion standards for health AI. IEEE Commun Stand Mag. 2020;4(3):64–69. https://doi.org/10.1109/ MCOMSTD.001.2000006. 8. Rajkomar, A., Dean, J., Kohane, I. Machine learning in medicine. N Engl J Med. 2019;380(14):1347–1358. https://doi.org/10.1056/NEJMra1814259. 9. Wenzel, M.A. and Wiegand, T. Towards international standards for the evaluation of artificial intelligence for health. In 2019 ITU kaleidoscope: ICT for health: Networks, standards and innovation (ITU K), Atlanta, GA, USA, 04–06 December 2019; IEEE Xplore: Pisca- teville, New Jersey, USA. 10. European Parliament. EU AI Act: First regulation on ar- tificial intelligence 2013. Available online: https://www. europarl.europa.eu/topics/en/article/20230601STO93804/ eu-ai-act-first-regulation-on-artificial-intelligence. 11. European Medicines Agency (EMA). Artificial intelligence workplan to guide use of AI in medicines regulation. Available online: https://www.ema.europa.eu/en/ news/artificial-intelligence-workplan-guide-use-ai- medicines-regulation. 12. Muehlematter, U.J., Daniore, P., Vokinger, K.N. Approval of artificial intelligence and machine learning-based medical devices in the USA and Europe (2015–20): A compara- tive analysis. Lancet Digit Health. 2021;3(3):e195–e203. https://doi.org/10.1016/S2589-7500(20)30292-2. 13. Sharma, K. and Manchikanti, P. AI-based medical de- vices and regulations: A cross-country perspective. In Artificial Intelligence in Drug Development. Springer Nature: Berlin, Germany; 2024; pp. 67–115. https:// doi.org/10.1007/978-981-97-2954-8_3. 14. Importance of compliance with regulations in the pharmaceutical industry. Available online: https:// doi.org/10.2139/SSRN.4679812. 15. EU reach political agreement on AI act – Key areas of impact for life sciences companies. Available online: https:// www.sidley.com/en/insights/newsupdates/2023/12/ eu-reach-political-agreement-on-ai-act. 16. European Medicines Agency. Medical devices. Available online: https://www.ema.europa.eu/en/ human-regulatory-overview/medical-devices. 17. Lexology. EU EMA proposes risk-based ap- proach to AI in pharma lifecycle. Available on- line: https://www.lexology.com/library/detail. aspx?g=2debbf90-5595-412d-b05d-744a5f116c27. https://www.scilife.io/blog/artificial-intelligence-regulations https://www.scilife.io/blog/artificial-intelligence-regulations https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-software-medical-device https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-software-medical-device https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-software-medical-device https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-software-medical-device https://doi.org/10.1038/s41746-023-00992-8 https://doi.org/10.1038/s41746-023-00992-8 https://www.pda.org/pda-letter-portal/home/full-article/a-comprehensive-review-of-regulatory-intelligence-exploring-tools-and-program-maturities https://www.pda.org/pda-letter-portal/home/full-article/a-comprehensive-review-of-regulatory-intelligence-exploring-tools-and-program-maturities https://www.pda.org/pda-letter-portal/home/full-article/a-comprehensive-review-of-regulatory-intelligence-exploring-tools-and-program-maturities https://www.pda.org/pda-letter-portal/home/full-article/a-comprehensive-review-of-regulatory-intelligence-exploring-tools-and-program-maturities http://ijcrt.org/viewfull.php?&p_id=IJCRT2304232 http://ijcrt.org/viewfull.php?&p_id=IJCRT2304232 https://doi.org/10.1007/s43441-024-00628-3 https://doi.org/10.1007/s43441-024-00628-3 https://doi.org/10.1109/MCOMSTD.001.2000006 https://doi.org/10.1109/MCOMSTD.001.2000006 https://doi.org/10.1056/NEJMra1814259. https://www.europarl.europa.eu/topics/en/article/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence https://www.europarl.europa.eu/topics/en/article/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence https://www.europarl.europa.eu/topics/en/article/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence https://www.ema.europa.eu/en/news/artificial-intelligence-workplan-guide-use-ai-medicines-regulation https://www.ema.europa.eu/en/news/artificial-intelligence-workplan-guide-use-ai-medicines-regulation https://www.ema.europa.eu/en/news/artificial-intelligence-workplan-guide-use-ai-medicines-regulation https://doi.org/10.1016/S2589-7500(20)30292-2 https://doi.org/10.1007/978-981-97-2954-8_3 https://doi.org/10.1007/978-981-97-2954-8_3 https://doi.org/10.2139/SSRN.4679812 https://doi.org/10.2139/SSRN.4679812 https://www.sidley.com/en/insights/newsupdates/2023/12/eu-reach-political-agreement-on-ai-act https://www.sidley.com/en/insights/newsupdates/2023/12/eu-reach-political-agreement-on-ai-act https://www.sidley.com/en/insights/newsupdates/2023/12/eu-reach-political-agreement-on-ai-act https://www.ema.europa.eu/en/human-regulatory-overview/medical-devices https://www.ema.europa.eu/en/human-regulatory-overview/medical-devices https://www.lexology.com/library/detail.aspx?g=2debbf90-5595-412d-b05d-744a5f116c27 https://www.lexology.com/library/detail.aspx?g=2debbf90-5595-412d-b05d-744a5f116c27 23 J Global Clinical Engineering Vol.7 Issue 2: 2025 Gaur, Akram, Singh, Chauhan: Artificial Intelligence-Driven Insights for Regulatory Intelligence in Medical Devices: Evaluating EMA, FDA, and CDSCO Frameworks 18. European Medicines Agency (EMA). Reflection paper on the use of artificial intelligence in the lifecycle of medicines. Available online: https://www.ema.europa.eu/en/news/ reflection-paper-use-artificial-intelligence-lifecycle-medicines. 19. European Medicines Agency (EMA). Reflection paper on the use of artificial intelligence (AI) in the medici- nal product lifecycle. Available online: https://www. ema.europa.eu/en/documents/scientific-guideline/ reflection-paper-use-artificial-intelligence-ai-medicinal- product-lifecycle_en.pdf. 20. Benjamens, S., Dhunnoo, P., Meskó, B. The state of artificial intelligence-based FDA-approved medical devices and algorithms: An online database. NPJ Digit Med. 2020;3:118. https://doi.org/10.1038/ s41746-020-00324-0. 21. Joshi, G., Jain, A., Araveeti, S.R., et al. FDA-approved artificial intelligence and machine learning (AI/ML)- enabled medical devices: An updated landscape. Electron. 2024;13(3):498. https://doi.org/10.3390/ electronics13030498. 22. Food and Drug Administration (FDA). Proposed regu- latory framework for modifications to artificial intel- ligence/machine learning (AI/ML)-based software as a medical device (SaMD). Available online: https:// www.fda.gov/files/medical%20devices/published/ US-FDA-Artificial-Intelligence-and-Machine-Learning- Discussion-Paper.pdf. 23. Food and Drug Administration (FDA). Marketing sub- mission recommendations for a predetermined change control plan for artificial intelligence/machine learning (AI/ML)-enabled device software functions. Available online: https://www.fda.gov/regulatory-information/ search-fda-guidance-documents/marketing-submission- recommendations-predetermined-change-control- plan-artificial. 24. Food and Drug Administration (FDA). FDA releases artificial intelligence/machine learning action plan. Available online: https://www.fda.gov/news-events/ press-announcements/fda-releases-artificial-intelli- gencemachine-learning-action-plan. 25. Food and Drug Administration (FDA). FDA action plan for AI/ML in SaMD (software as a medical device). Available online: https://starfishmedical.com/blog/ fda-action-plan-for-ai-ml-in-samd-software-as-a- medical-device/. 26. Akin. FDA releases action plan for artificial intelligence/ machine learning-enabled software as a medical device. Available online: https://www.akingump.com/en/insights/ alerts/fda-releases-action-plan-for-artificial-intelligencema- chine-learning-enabled-software-as-a-medical-device-provi- ding-another-iterative-step-forward-on-the-long-road-ahead. 27. Zhu, S., Gilbert, M., Chetty, I., et al. The 2021 landscape of FDA-approved artificial intelligence/machine learning-enabled medical devices: An analysis of the characteristics and intended use. Int J Med In- form. 2022;165:104828. https://doi.org/10.1016/J. IJMEDINF.2022.104828. 28. Food and Drug Administration (FDA). Artificial in- telligence and machine learning (AI/ML)-enabled medical devices. Available online: https://www.fda. gov/medical-devices/software-medical-device-samd/ artificial-intelligence-and-machine-learning-aiml- enabled-medical-devices. 29. Ropes & Gray. Getting smarter: FDA publishes draft guidance on predetermined change control plans for artificial intelligence/machine learning (AI/ML) devices. Available online: https://www.ropesgray. com/en/insights/alerts/2023/05/getting-smarter- fda-publishes-draft-guidance-on-predetermined- change-control-plans-for-ai-ml-devices. 30. LeCun, Y., Bengio, Y., Hinton, G. Deep learning. Na- ture. 2015;521:436–444. https://doi.org/10.1038/ NATURE14539. 31. Gerke, S., Babic, B., Evgeniou, T., et al. The need for a system view to regulate artificial intelligence/ma- chine learning-based software as medical device. NPJ Digit Med. 2020;3(1):53. https://doi.org/10.1038/ S41746-020-0262-2. https://www.ema.europa.eu/en/news/reflection-paper-use-artificial-intelligence-lifecycle-medicines https://www.ema.europa.eu/en/news/reflection-paper-use-artificial-intelligence-lifecycle-medicines https://www.ema.europa.eu/en/documents/scientific-guideline/reflection-paper-use-artificial-intelligence-ai-medicinal-product-lifecycle_en.pdf https://www.ema.europa.eu/en/documents/scientific-guideline/reflection-paper-use-artificial-intelligence-ai-medicinal-product-lifecycle_en.pdf https://www.ema.europa.eu/en/documents/scientific-guideline/reflection-paper-use-artificial-intelligence-ai-medicinal-product-lifecycle_en.pdf https://www.ema.europa.eu/en/documents/scientific-guideline/reflection-paper-use-artificial-intelligence-ai-medicinal-product-lifecycle_en.pdf https://doi.org/10.1038/s41746-020-00324-0 https://doi.org/10.1038/s41746-020-00324-0 https://doi.org/10.3390/electronics13030498 https://doi.org/10.3390/electronics13030498 https://www.fda.gov/files/medical%20devices/published/US-FDA-Artificial-Intelligence-and-Machine-Learning-Discussion-Paper.pdf https://www.fda.gov/files/medical%20devices/published/US-FDA-Artificial-Intelligence-and-Machine-Learning-Discussion-Paper.pdf https://www.fda.gov/files/medical%20devices/published/US-FDA-Artificial-Intelligence-and-Machine-Learning-Discussion-Paper.pdf https://www.fda.gov/files/medical%20devices/published/US-FDA-Artificial-Intelligence-and-Machine-Learning-Discussion-Paper.pdf https://www.fda.gov/regulatory-information/search-fda-guidance-documents/marketing-submission-recommendations-predetermined-change-control-plan-artificial https://www.fda.gov/regulatory-information/search-fda-guidance-documents/marketing-submission-recommendations-predetermined-change-control-plan-artificial https://www.fda.gov/regulatory-information/search-fda-guidance-documents/marketing-submission-recommendations-predetermined-change-control-plan-artificial https://www.fda.gov/regulatory-information/search-fda-guidance-documents/marketing-submission-recommendations-predetermined-change-control-plan-artificial https://www.fda.gov/news-events/press-announcements/fda-releases-artificial-intelligencemachine-learning-action-plan https://www.fda.gov/news-events/press-announcements/fda-releases-artificial-intelligencemachine-learning-action-plan https://www.fda.gov/news-events/press-announcements/fda-releases-artificial-intelligencemachine-learning-action-plan https://starfishmedical.com/blog/fda-action-plan-for-ai-ml-in-samd-software-as-a-medical-device/ https://starfishmedical.com/blog/fda-action-plan-for-ai-ml-in-samd-software-as-a-medical-device/ https://starfishmedical.com/blog/fda-action-plan-for-ai-ml-in-samd-software-as-a-medical-device/ https://www.akingump.com/en/insights/alerts/fda-releases-action-plan-for-artificial-intelligencemachine-learning-enabled-software-as-a-medical-device-providing-another-iterative-step-forward-on-the-long-road-ahead https://www.akingump.com/en/insights/alerts/fda-releases-action-plan-for-artificial-intelligencemachine-learning-enabled-software-as-a-medical-device-providing-another-iterative-step-forward-on-the-long-road-ahead https://www.akingump.com/en/insights/alerts/fda-releases-action-plan-for-artificial-intelligencemachine-learning-enabled-software-as-a-medical-device-providing-another-iterative-step-forward-on-the-long-road-ahead https://www.akingump.com/en/insights/alerts/fda-releases-action-plan-for-artificial-intelligencemachine-learning-enabled-software-as-a-medical-device-providing-another-iterative-step-forward-on-the-long-road-ahead https://doi.org/10.1016/J.IJMEDINF.2022.104828 https://doi.org/10.1016/J.IJMEDINF.2022.104828 https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-aiml-enabled-medical-devices https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-aiml-enabled-medical-devices https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-aiml-enabled-medical-devices https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-aiml-enabled-medical-devices https://www.ropesgray.com/en/insights/alerts/2023/05/getting-smarter-fda-publishes-draft-guidance-on-predetermined-change-control-plans-for-ai-ml-devices https://www.ropesgray.com/en/insights/alerts/2023/05/getting-smarter-fda-publishes-draft-guidance-on-predetermined-change-control-plans-for-ai-ml-devices https://www.ropesgray.com/en/insights/alerts/2023/05/getting-smarter-fda-publishes-draft-guidance-on-predetermined-change-control-plans-for-ai-ml-devices https://www.ropesgray.com/en/insights/alerts/2023/05/getting-smarter-fda-publishes-draft-guidance-on-predetermined-change-control-plans-for-ai-ml-devices https://doi.org/10.1038/NATURE14539 https://doi.org/10.1038/NATURE14539 https://doi.org/10.1038/S41746-020-0262-2 https://doi.org/10.1038/S41746-020-0262-2 Gaur, Akram, Singh, Chauhan: Artificial Intelligence-Driven Insights for Regulatory Intelligence in Medical Devices: Evaluating EMA, FDA, and CDSCO Frameworks J Global Clinical Engineering Vol.7 Issue 2: 2025 24 32. Bharadwaj, M.S. Regulatory insights in digital health. In Multi-sector analysis of the digital healthcare industry, Chatterjee, L., Gani, N., IGI Global Scientific Publishing: Hershey, PA; USA, 2024; pp.164–197. 33. DDReg Pharma. Regulatory guidelines for software and AI medical devices. Available online: https://resource. ddregpharma.com/blogs/regulatory-guidelines-for- software-and-artificial-intelligence-as-a-medical-device/. 34. Central Drugs Standard Control Organization (CD- SCO) Directorate General of Health Services Ministry of Health and Family Welfare Government of India. Available online: https://cdsco.gov.in/opencms/re- sources/UploadCDSCOWeb/2018/UploadTenderFile/ Corrigendum_to_CDSCO_EoI_SSP_01st_Dec.pdf. 35. Freyr Solutions. Regulation of software as medical device (SaMD) in India. Available online: https://www.freyrsolutions.com/blog/ regulation-of-software-as-medical-device-samd-in-india. 36. International Medical Device Regulators Forum (IMDRF). Essential Principles of Safety and Performance of Medical Devices and IVD Medical Devices. Available online: https:// www.imdrf.org/sites/default/files/2024-04/IMDRF%20 GRRP%20WG%20N47%20%28Edition%202%29.pdf. 37. International Medical Device Regulators Forum (IMDRF). Medical device regulatory review report: Guidance regarding information to be included. Avail- able online: https://www.imdrf.org/sites/default/ files/2024-04/IMDRF%20GRRP%20WG%20N71%20 %28Edition%202%29_0.pdf. 38. International Medical Device Regulators Forum (IM- DRF). Principles of labeling for medical devices and IVD medical devices. Available online: https://www. imdrf.org/sites/default/files/2024-04/IMDRF%20 GRRP%20WG%20N52%20%28Edition%202%29.pdf. 39. Rajpurkar, P., Chen, E., Banerjee, O., et al. AI in health and medicine. Nat Med. 2022;28:31–38. https://doi. org/10.1038/S41591-021-01614-0. 40. Comparative analysis of artificial intelligence on medi- cal device regulations and legislation in US and EU. Available online: https://www.theseus.fi/bitstream/ handle/10024/506136/Nawar_Bushra_2021_Com- parative%20Analysis%20of%20Artificial%20Intelli- gence%20on%20Medical%20Device%20Regulations%20 and%20Legislation%20in%20US%20and%20EU. pdf?sequence=2. 41. How the challenge of regulating AI in healthcare is escalating. Available online: https://www.ey.com/en_gl/insights/law/ how-the-challenge-of-regulating-ai-in-healthcare-is-escalating. 42. Onitiu, D., Wachter, S., Mittelstadt, B. How AI challenges the medical device regulation: Patient safety, benefits, and intended uses. J Law Biosci. 2024;lsae007. https:// doi.org/10.1093/jlb/lsae007. 43. Raji, I.D., Kumar, I.E., Horowitz, A., et al. The fallacy of AI functionality. ACM Int Conf Proc. 2022:959–972 (Published online June 21). https://doi.org/10.1145/3531146.3533158. 44. Matheny, M.E., Whicher, D., Thadaney Israni, S. Artificial intelligence in health care: A report from the National Academy of Medicine. JAMA. 2020;323(6):509–510. https://doi.org/10.1001/jama.2019.21579. 45. Khan, B., Fatima, H., Qureshi, A., et al. Drawbacks of artificial intelligence and their potential solutions in the healthcare sector. Biomed Mat Devices. 2023;1:731–738. https://doi.org/10.1007/S44174-023-00063-2. 46. Henry, E. and Thiel, S. Using existing regulatory frameworks to apply effective design controls to AI/ML in medical devices. Biomed Instrum Technol. 2022;56(4):114–118. https://doi.org/10.2345/0899-8205-56.4.114. 47. Palaniappan, K., Lin, E.Y.T, Vogel, S. Global regulatory frameworks for the use of artificial intelligence (AI) in the healthcare services sector. Healthcare. 2024;12(5):562. https://doi.org/10.3390/HEALTHCARE12050562. 48. Center for Devices and Radiological Health. (2022, September 26). Digital Health Software Precertifica- tion (Pre-CERT) pilot program. U.S. Food And Drug Administration. Available online: https://www.fda. gov/medical-devices/digital-health-center-excellence/ digital-health-software-precertification-pre-cert-pilot-program. 49. Gerke, S., Minssen, T., Cohen, G. Ethical and legal chal- lenges of artificial intelligence-driven healthcare. Artificial Intelligence in Healthcare (1st edition); Bohr, A. and Memarzadeh, K. (eds.);2020:295–336. https:// doi.org/10.1016/B978-0-12-818438-7.00012-5. https://resource.ddregpharma.com/blogs/regulatory-guidelines-for-software-and-artificial-intelligence-as-a-medical-device/ https://resource.ddregpharma.com/blogs/regulatory-guidelines-for-software-and-artificial-intelligence-as-a-medical-device/ https://resource.ddregpharma.com/blogs/regulatory-guidelines-for-software-and-artificial-intelligence-as-a-medical-device/ https://cdsco.gov.in/opencms/resources/UploadCDSCOWeb/2018/UploadTenderFile/Corrigendum_to_CDSCO_EoI_SSP_01st_Dec.pdf https://cdsco.gov.in/opencms/resources/UploadCDSCOWeb/2018/UploadTenderFile/Corrigendum_to_CDSCO_EoI_SSP_01st_Dec.pdf https://cdsco.gov.in/opencms/resources/UploadCDSCOWeb/2018/UploadTenderFile/Corrigendum_to_CDSCO_EoI_SSP_01st_Dec.pdf https://www.freyrsolutions.com/blog/regulation-of-software-as-medical-device-samd-in-india https://www.freyrsolutions.com/blog/regulation-of-software-as-medical-device-samd-in-india https://www.imdrf.org/sites/default/files/2024-04/IMDRF%20GRRP%20WG%20N47%20%28Edition%202%29.pdf https://www.imdrf.org/sites/default/files/2024-04/IMDRF%20GRRP%20WG%20N47%20%28Edition%202%29.pdf https://www.imdrf.org/sites/default/files/2024-04/IMDRF%20GRRP%20WG%20N47%20%28Edition%202%29.pdf https://www.imdrf.org/sites/default/files/2024-04/IMDRF%20GRRP%20WG%20N71%20%28Edition%202%29_0.pdf https://www.imdrf.org/sites/default/files/2024-04/IMDRF%20GRRP%20WG%20N71%20%28Edition%202%29_0.pdf https://www.imdrf.org/sites/default/files/2024-04/IMDRF%20GRRP%20WG%20N71%20%28Edition%202%29_0.pdf https://www.imdrf.org/sites/default/files/2024-04/IMDRF%20GRRP%20WG%20N52%20%28Edition%202%29.pdf https://www.imdrf.org/sites/default/files/2024-04/IMDRF%20GRRP%20WG%20N52%20%28Edition%202%29.pdf https://www.imdrf.org/sites/default/files/2024-04/IMDRF%20GRRP%20WG%20N52%20%28Edition%202%29.pdf https://doi.org/10.1038/S41591-021-01614-0 https://doi.org/10.1038/S41591-021-01614-0 https://www.theseus.fi/bitstream/handle/10024/506136/Nawar_Bushra_2021_Comparative%20Analysis%20of%20Artificial%20Intelligence%20on%20Medical%20Device%20Regulations%20and%20Legislation%20in%20US%20and%20EU.pdf?sequence=2 https://www.theseus.fi/bitstream/handle/10024/506136/Nawar_Bushra_2021_Comparative%20Analysis%20of%20Artificial%20Intelligence%20on%20Medical%20Device%20Regulations%20and%20Legislation%20in%20US%20and%20EU.pdf?sequence=2 https://www.theseus.fi/bitstream/handle/10024/506136/Nawar_Bushra_2021_Comparative%20Analysis%20of%20Artificial%20Intelligence%20on%20Medical%20Device%20Regulations%20and%20Legislation%20in%20US%20and%20EU.pdf?sequence=2 https://www.theseus.fi/bitstream/handle/10024/506136/Nawar_Bushra_2021_Comparative%20Analysis%20of%20Artificial%20Intelligence%20on%20Medical%20Device%20Regulations%20and%20Legislation%20in%20US%20and%20EU.pdf?sequence=2 https://www.theseus.fi/bitstream/handle/10024/506136/Nawar_Bushra_2021_Comparative%20Analysis%20of%20Artificial%20Intelligence%20on%20Medical%20Device%20Regulations%20and%20Legislation%20in%20US%20and%20EU.pdf?sequence=2 https://www.theseus.fi/bitstream/handle/10024/506136/Nawar_Bushra_2021_Comparative%20Analysis%20of%20Artificial%20Intelligence%20on%20Medical%20Device%20Regulations%20and%20Legislation%20in%20US%20and%20EU.pdf?sequence=2 https://www.ey.com/en_gl/insights/law/how-the-challenge-of-regulating-ai-in-healthcare-is-escalating https://www.ey.com/en_gl/insights/law/how-the-challenge-of-regulating-ai-in-healthcare-is-escalating https://doi.org/10.1093/jlb/lsae007 https://doi.org/10.1093/jlb/lsae007 https://doi.org/10.1145/3531146.3533158 https://doi.org/10.1001/jama.2019.21579 https://doi.org/10.1007/S44174-023-00063-2 https://doi.org/10.2345/0899-8205-56.4.114 https://doi.org/10.3390/HEALTHCARE12050562 https://www.fda.gov/medical-devices/digital-health-center-excellence/digital-health-software-precertification-pre-cert-pilot-program https://www.fda.gov/medical-devices/digital-health-center-excellence/digital-health-software-precertification-pre-cert-pilot-program https://www.fda.gov/medical-devices/digital-health-center-excellence/digital-health-software-precertification-pre-cert-pilot-program https://doi.org/10.1016/B978-0-12-818438-7.00012-5 https://doi.org/10.1016/B978-0-12-818438-7.00012-5