









































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

