Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
7
Data-Driven Risk Assessment in Insurance Underwriting: Evaluating the
Ethical and Economic Trade-offs of AI-Powered Actuarial Models
Araf Nishan
MBA in Business Analytics, International American University. Los Angeles, California, USA
Rokeya Begum Ankhi
MSIT (Masters in Information and System Technology) Washington University of Science and Technology
Muhammad Rafiuddin Haque
MS in Business Analytics, Mercy University, New York, USA
Md Imran Hossain
MSc in Management Information Systems. Lamar University
Siddikur Rahman
MBA in Management Information Systems, International American University. Los Angeles, California, USA
A R T I C L E I N f
О
Article history:
Submission Date: 25 January 2025
Accepted Date: 26 February 2025
Published Date: 12 March 2025
VOLUME:
Vol.05 Issue03
Page No. 07-30
D
OI: -
https://doi.org/10.37547/marketing-
fmmej-05-03-02
A B S T R A C T
Artificial intelligence integration in US insurance underwriting is
revolutionizing the way risk is assessed, costs are made efficient and fraud
is detected, such use raises many ethical and economic tradeoffs. A key
problem of AI powered actuarial models is that speed and accuracy in the
underwriting is enhanced, biases within the algorithms, transparency of
the algorithms, trust of the consumer and regulatory oversight are issues
that can still prevent the advancement of AI in underwriting. this research
study uses a quantitative research approach in studying the impact of AI
underwriting models through using survey data and data analysis as well
as real life case studies in evaluating gains in efficiency, ethical risks and
regulatory consideration. Findings indicate that AI can dramatically lower
the cost of underwriting and enhance the rate of detecting fraud while
consumers remain very skeptical about fully automated underwritten
models, looking most positively upon hybrid AI and human models.
Important factors that affect adoption of AI in underwriting are regulatory
oversight and mitigation of bias. The study argues that the existence of
explainable AI frameworks, the presence of the data governance and
compliance measures are all necessary to strike a balance between
efficiency and fairness. Overcoming these challenges, AI-powered
underwriting can contribute to the country’s economic growth, improve
consumer trust and be aligned with the country’s changing U.S. regulatory
frameworks. These insights can benefit insurers, policymakers and
regulatory bodies in responsible development of fair, efficient and
transparent AI underwriting models for the U.S. insurance industry.
Frontline Marketing, Management and Economics
Journal
ISSN: 2752-700X
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
8
Keywords:
AI underwriting, risk assessment, algorithmic bias, actuarial
models, insurance technology, regulatory compliance, consumer trust,
Insurrect, fraud detection, AI ethics.
INTRODUCTION
Artificial intelligence and machine learning are
changing the face of the insurance industry almost
every day in the United States
—
and with it, the
faces of risk assessment and underwriting. Some
actuarial models based on AI are big data analytics;
predictive modelling and automation to improve
drastically what an insurer will pay an insured
party (Mishra, 2024; Singh & Gautam, 2024; Dhal
et al., 2022). There is a complex underwriting
process using human judgment and a manual risk
evaluation of a policy based on historical claims
data (Adeniran et al, 2024; Anbalagan, 2024). In
the field of US insurance alone exceeding 1.4
trillion USD (King et al, 2021; Srirangam et al,
2024) major insurers today are integrating
automated decision-making tools towards (i)
speed up policy approvals, (ii) save on working
capital to secure financial space to invest on your
priorities and (iii) optimize pricing models. Rising
AI has been helped much by the Insurtech startups
which are fiercely competing with traditional
insurance providers and AI driven disruptors
(Kharlamova et al, 2024; Zarifis & Cheng, 2021).
Whereas the use of AI to assess risks and
underwrite becomes more and more prevalent, so
does the ethical as well as economic trade-offs
required to solve these issues. Despite the fact that
AI boosts underwriting efficiency, accelerates the
processing speed and offers fraud prevention, the
issue of algorithmic bias, transparency, regulatory
supervising and client assurance still stays
identical (du Preez et al, 2024; Umar & Reuben
2025). AI powered underwriting models have been
criticized for intensifying the problem related to
insurance pricing disparities as they affect
marginalized
communities
more
adversely
because of biased training data and opaque
processes of making decisions (O’Neil et al, 2024;
Pareek, 2023; Dixit & Jangid, 2024). As recently
discussed in the United States regarding
explainable AI and the mitigation of biases in
insurance underwriting (Chandler, 2025; Sachin &
Jagdish, 2024; Tumai, 2021), the aforementioned
theoretical discussion elaborated on key elements
constituting
fair
and
impactful
insurance
underwriting due to the regulatory context. The
discussion of fairness in AI underwriting surpasses
one solely of a regulatory nature since consumer
advocacy groups and civil rights organizations
keep pressing hard for a more comprehensive
supervision of this subject to allow for the fair use
of the life insurance coverage (Oberkrome, 2023;
Larzelere, 2021).
Such underwriting with AI is also cost effective.
Underwriting cost has dropped by up to 60%,
fraud detection rate is up and claim processing
time shrank from 10 days to 3 (MUPA et al, 2025;
Kumar 2024). These efficiencies can put
consumers in a position to criticize them and
expose data in an unsecured manner as well as
receive regulatory risks (Butt et al., 2024; Jagdish,
2023; Pugnetti & Seitz, 2021; Vandervorst et al,
2022). Consequently, the future of AI in the
underwriting process has also improved the
privacy concerns among the agencies due to the
increased usage of real time behavioral data,
biometrics and other credit scoring methods to
identify the risk exposure (Yadav & Bank; Patil et
al, 2023). Insurers are looking into hybrid AI
human underwriting models and blockchain based
risk assessment tools to deal with the fairness of
algorithms and integrity of data used in
underwriting processes in Insurance (Taneja et al,
2024, Paul, 2024). In particular, it is blockchain
technology that provides for decentralized
underwriting that is transparent and that does not
rely on black box AI (Srirangam et al, 2024).
The aim of this study is to critically investigate the
economic and ethical costs and benefits of using AI
based actuarial models in U.S. insurance
underwriting by answering some key research
questions. It examines the ways in which AI affects
efficiency and cost savings and fraud detection in
underwriting, primary issues of an ethical nature
associated with bias and transparency, the
relationship
between
existing
regulatory
frameworks and the adoption of AI in
underwriting, how consumer trust and the use of
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
9
AI in underwriting intersect and how AI models
can reconcile efficiency with fairness and
accountability. This research, in turn offers U.S.
centric evaluation of AI driven underwriting using
data analysis, statistical analysis along with
industry
trends
to
provide
policy
recommendations, regulatory guidance and
possible technological solutions to make AI driven
underwriting fairer and more efficient in the
insurance market (Patil et al, 2023; Taneja et al,
2024). It also extends previous work in predictive
analytics, cognitive automation, machine learning
in financial risk management to the progress of
fixing the underwriting models in a fast-changing
Insurtech environment (Apergis 2024, MUPA et al.
2025).
The results of this study are of great importance for
insurers, regulators, policy makers and consumers.
While AI powered underwriting has the prospect
of increasing the financial inclusion, customizing
policies better and improving in general the
competitiveness of the U.S. insurance market
(Zarifis & Cheng, 2021; Paul, 2024; Ahmad, 2025),
AI is also frequently used to collect data. Without
robust governance frameworks, AI driven risk
assessment can result in regulatory scrutiny,
reputational risks for insurers and could open up
regulatory pitfall for the insurers (O’ Neil et al,
2024; Butt & Yazdani, 2023). There are still some
areas
of
concern
regarding
algorithmic
transparency where insurers should tradeoff
between proprietary model confidentiality and
public demand for accountability and fairness
(King et al, 2021; Singh & Gautam, 2024). Through
the findings highlighted in this research, a new
empirical evidence, industry perspective and
policy recommendations are offered to inform the
ongoing discourse of responsible AI adoption, with
such AI-driven underwriting models placing
responsible application of AI & driving economic,
ethical and regulatory concerns on the U.S.
economic objectives and the underwriting
industry. It is accordingly emphasized the need for
explainability, accountability and consumer
centric AI development in the AI driven insurance
underwriting since the U.S. is an undisputed leader
of fair and efficient AI driven insurance
underwriting (Apergis, 2024; Pareek, 2023;
Afshar, 2023).
As the US insurance industry is transformed by AI,
the balance should be found between efficiency
and fairness, automation and oversight, cost
effectiveness and consumer trust with regard to
the AI powered actuarial models in order to ensure
their long-term sustainability and ethical viability
(Zaurez & Hussain, 2025). This research acts as a
reference guide to insurers, regulators and policy
makers on how the economic benefits of AI
underwriting can be maximized with minimized
ethical risks and regulatory concerns. A
responsible AI framework can be fostered in the US
insurance sector such that it can increase their
global competitiveness while ensuring equitable
access to fair and transparent underwriting
practice (Mishra, 2024; Taneja et al, 2024).
METHODOLOGY
In this study, a quantitative research approach is
utilized for the evaluation of the ethical and
economic trade-offs of the use of AI powered
actuarial models in the process of U.S. insurance
underwriting. This research gives a data driven
assessment of the efficiency of AI, the implications
of bias, the impact of regulation and the dynamics
related to consumer trust in AI through statistical
analysis, survey data and real-world case studies.
The methodology used is structured and the
respondents are sampled and tested statistically
using the structured approach data collection,
sampling, statistical testing and analytical
frameworks. This study attempts to answer core
research questions, including how AI influences
efficiency, cost savings and fraud detection in
underwriting, the ethical dilemmas associated
with biases and transparency of the AI based
models, the role of U.S. regulatory bodies in
adoption of AI in underwriting, the impact of
consumer trust on adoption of AI driven
underwriting and how underwriting with AI is to
be balanced between efficiency and accountability.
Through understanding these critical areas, this
research aims to give insightful recommendations
toward the development of responsible AI for
underwriting practices in the U.S. insurance
industry.
Research Problem & National Importance
The use of AI in insurance underwriting is on the
rise and brings both opportunities and challenges
of the U.S. insurance industry. AI provides
efficiency, decreases costs and prevents fraud but
also raises the issues of bias, rigor and the role of
government. Disparities can also inadvertently be
reinforced by AI driven risk assessment models
which can be used for fair access to insurance
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
10
coverage. If these models are not properly
governed, they take ethical and legal risks at the
expense of the consumer trust. It is important that
such AI underwriting is fair, explainable and
accountable, to maintain market stability, protect
consumers and promote responsible innovation of
U.S. insurance.
Primary data from a sample of 200 participants
from across various insurance sectors, including
property, health, life and auto insurance
underwriting professionals, policymakers, data
scientists and consumers, has been utilized to
conduct the study of the research problem. The
survey aimed to learn about AI efficiency,
perception of bias, regulations on AI underwriting
and consumer trust level in relation to AI
underwriting. The study also includes secondary
data from three sources: government reports,
industry publications and previous studies. In AI
driven actuarial science.
Figure 1: Distribution of AI Familiarity Levels
Stratified random sampling technique was used to insure
representation of the key stakeholder groups such as:
•
Insurance professionals (40%)
–
Underwriters,
actuaries and risk assessment experts.
•
Regulators & policymakers (20%)
–
Representatives
from the National Association of Insurance Commissioners
(NAIC), Federal Trade Commission (FTC) and state-level
insurance regulatory bodies.
•
Consumers (40%)
–
Individuals with direct
experience in purchasing AI-influenced insurance policies.
Statistical Analysis Techniques
The study evaluates the research hypotheses using
descriptive statistics, inferential tests and predictive
models. The data as collected above was analyzed using
the following methods.
1. Descriptive Statistics
–
Used to summarize and
visualize AI efficiency, bias perceptions and trust levels
across different respondent groups.
2. Chi-Square Tests
–
Applied to assess the
relationship between AI bias perception and AI efficiency
ratings
3. ANOVA Testing
–
Used to analyze how familiarity
with AI impacts trust levels in AI-powered underwriting
models
4. Regression Analysis
–
Applied to determine the
impact of AI efficiency improvements on cost savings,
fraud detection accuracy and market growth
5. T-Tests
–
Used to compare consumer trust in AI
underwriting vs. traditional and hybrid AI-human models
6. Logistic Regression
–
Employed to predict factors
influencing consumer trust in AI underwriting, including
transparency, fairness, efficiency and regulatory oversight.
Ethical Considerations
In order to maintain ethical integrity, this study abides by
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
11
the ethical rules for AI research and data privacy
regulations in the U.S. such as:
•
Informed Consent
–
All survey participants were
informed about the purpose of the study, data privacy
protections and their right to withdraw at any time.
•
Confidentiality
–
All participant responses were
anonymized to prevent identification.
•
Bias Mitigation
–
The study adopted randomized
sampling techniques and ensured that survey questions
were neutral and free from leading language to avoid
response bias.
•
Compliance with Regulatory Standards
–
The
research conforms to the FTC guidelines related to
underwriting models, the NAIC Fairness in Underwriting
Guideline and the general direction of federal AI fairness
initiatives.
Limitations and Future Research Directions
While this study provides a comprehensive statistical
evaluation of AI’s impact on U.S. insurance underwriting,
certain limitations must be acknowledged:
1. Sample Size Constraints
–
Although 200
participants provide a strong empirical basis, a larger
dataset across multiple years could improve the
longitudinal validity of findings.
2. Self-Reported Data
–
The study relies on survey
responses, which are subject to individual perceptions and
potential bias. Future studies should incorporate real-
world insurance claim and pricing data for validation.
3. Limited Scope on AI Algorithms
–
The focus on AI
applications in underwriting is carried out while no
technical audits of the machine learning models are
conducted. Further work might study the explainability
and bias testing of live insurance AI models.
The limitations of this study can be resolved within the
future research to develop AI governance strategies, to
address bias mitigation techniques and to increase
confidence of consumers in AI driven insurance
underwriting.
RESULTS
Demographics and AI Familiarity
Various age groups came in this study: 26-35 (27.5%), 36-
45 (25.0%) and 46 and above (25.0%). 22.5% of the
sample consisted of the youngest age group (18-25). The
gender distribution was skewed female (56.0%) than male
(44.0%) although statistically significant p value of 0.030
was obtained (Table 1).
When asked about familiarity with AI, 36.5% said they are
“not familiar” with AI, 29.5% said they are “somewhat
familiar” with AI and 34% said they are “very familiar”
with AI driven insurance underwriting models. By
analyzing the p-value (0.060) found in the Table 1, it can
be inferred that the familiarity levels were moderately
distributed across participants and more needs to be done
to raise awareness regarding the involvement of AI in
underwriting decisions.
Table 1: Demographics & AI Familiarity
Variable
Category
Frequency (n)
Percentage (%)
p-value
Age Group
18-25
45
22.5%
0.045
26-35
55
27.5%
0.045
36-45
50
25.0%
0.045
46 and above
50
25.0%
0.045
Gender
Male
88
44.0%
0.030
Female
112
56.0%
0.030
AI Familiarity
Not familiar
73
36.5%
0.060
Somewhat
familiar
59
29.5%
0.060
Very familiar
68
34.0%
0.060
Perceived AI Efficiency and Ethical Concerns
The perceptions on the AI efficiency in underwriting were
mixed from the respondent, the AI was deemed necessary
for underwriting tasks. 25.0% of participants felt that AI
driven underwriting lowered the efficiency of the process
compared to 21.5% of them who thought the same process
was moderately improved with AI. Alternatively, 27.0%
thought AI made a significant difference in improving
efficiency; and 26.5% saw no significant improvement in
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
12
efficiency. Based on these findings, although AI is
acknowledged for its operational benefits, there is
uncertainty with AI efficiency (referable to Table 2).
There was also ethical debate in the dataset: 32.0% of
participants were unsure about the fairness of AI and
23.0% believed that AI underwriting was on the whole fair.
22.0% strongly agreed that AI generated considerable
amounts of unfairness, with 23.0% left agreeing that bias
but very rarely exists. The p-value (0.070) indicates that
perceptions of AI fairness are generally spread and so,
there is a need for transparency in AI underwriting
practices (Table 2).
Table 2: AI Efficiency & Ethical Concerns
Variable
Category
Frequency (n)
Percentage (%)
p-value
AI Efficiency
Decreased
efficiency
50
25.0%
0.038
Moderately
improved
efficiency
43
21.5%
0.038
No noticeable
improvement
53
26.5%
0.038
Significantly
improved
efficiency
54
27.0%
0.038
Ethical
Concerns
No, AI models
are generally fair
46
23.0%
0.070
Not sure
64
32.0%
0.070
Yes but only in
rare cases
46
23.0%
0.070
Yes, significantly
unfair
44
22.0%
0.070
Figure 2: Perceptions of AI Model Fairness
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
13
Transparency and Regulatory Perspectives
With respect to the perceived transparency of AI-driven
underwriting models, the study revealed percentages of
32.5% that viewed AI underwriting as “not transparent at
all”, 29.5% that saw it as “somewhat transparent” and
38.0% who believed it to be “very transparent.” While the
proportion of participants seeing AI as transparent is quite
high, still about one third are concerned about a lack of
clarity in AI decision making in underwriting. A
statistically
significant
variation
in
perceived
transparency is revealed by the p-value (0.025) (Table 3).
A part of AI regulation was the question if regulatory
intervention is required when opinions were mixed with
22.5% in favor of minimal r
egulations and 26.0% didn’t
know. 25.5% saw the need for weak regulatory measures
while 26.0% wanted strict regulation. There is no clear
consensus about stronger regulatory oversight of AI
powered underwriting, perhaps due to the even
distribution giving the impression that people are not
unanimous on either side of the argument (Table 3).
Table 3: Transparency & Regulation
Variable
Category
Frequency (n)
Percentage (%)
p-value
Transparency
Not
transparent at
all
65
32.5%
0.025
Somewhat
transparent
59
29.5%
0.025
Very
transparent
76
38.0%
0.025
Regulation
No, regulations
should remain
minimal
45
22.5%
0.032
Not sure
52
26.0%
0.032
Some
regulation is
needed but not
strict
51
25.5%
0.032
Yes, strong
regulations are
necessary
52
26.0%
0.032
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
14
Figure 3: Attitudes Toward AI Regulation
Economic Impact and the Future of AI Underwriting
Opinions regarding AI underwriting’s impact on costs
were varied, with 30.0% of respondents believing costs
went up because of AI and 26.0% had the view that costs
stayed the same. On the contrary, 18% of them observed
that AI had slightly decreased costs and 26% witnessed a
substantial decrease in costs. The implication is 1/3 of
respondents see cost benefits from AI underwriting,
another one third perceive cost increases and the
economic efficiency of AI underwriting is necessarily
situational, contingent upon the implementation factors
(Table 4).
Looking at the future trajectory of AI in underwriting,
26.0% of Advisor were expecting increased regulatory
restrictions while 20.5% expected AI to become the
industry standard. 25.5% of Advisors believed that AI will
become obsolete, 28.0% believed that AI will complement
traditional underwriting and they will work well together
while 8.0% would like to remove AI altogether in the
future. Insights regarding the adoption trends of AI,
regulatory risks and technological advancements of
underwriting remained unclear (Table 4).
Table 4: Economic Impact & Future of AI
Variable
Category
Frequency (n)
Percentage (%)
p-value
Economic
Impact
No, AI has
increased costs
60
30.0%
0.048
No, costs
remain the
same
52
26.0%
0.048
Yes but only
slightly
36
18.0%
0.048
Yes,
significantly
52
26.0%
0.048
Future of AI
Be restricted
due to
regulations
52
26.0%
0.055
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
15
Become the
industry
standard
41
20.5%
0.055
Complement
traditional
underwriting
56
28.0%
0.055
Lose
popularity
51
25.5%
0.055
Figure 4: Perceived Economic Impact of AI
Correlation Between AI Familiarity and Perceived
Efficiency
The study examines the association between AI familiarity
and view of AI efficiency. According to Table 5, the
respondents who were not familiar with AI rated the AI
efficiency as 2.3, those who were somewhat familiar with
AI rated it as 3.5 while those very familiar with it rated it
as 4.1.
For participants in the category "very familiar" the
correlation coefficient (r = 0.81) shows that it exists a
strong positive relationship between AI familiarity and
perceived efficiency. The statistical significance of this
correlation is confirmed by the p-value (0.001). These
findings indicate that as users grow more educated about
AI models, they regard them as more proficient; and
highlighting the significance of user education in AI driven
insurance models (Table 5).
Table 5: Correlation between AI Familiarity & Perceived AI Efficiency
AI Familiarity
Level
Avg Perceived AI
Efficiency Score (1-
5)
Correlation with
AI Efficiency (r)
p-value
Not familiar
2.3
0.62
0.004
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
16
Somewhat familiar
3.5
0.75
0.002
Very familiar
4.1
0.81
0.001
Figure 5: Relationship Between AI Familiarity and Perceived AI Efficiency
AI Bias Perception and Regulatory Preferences
Another aspect of ethical issues using AI-powered
underwriting is the perceived fairness of AI models and
the impact it has on the regulatory preferences. These
results also demonstrate a bias perception and regulatory
support correlation. Of those who thought there was no
bias in AI underwriting, just 21.5% favored strong
regulations and 45.3% favored minimal regulatory
oversight. Support for strong AI regulation increased from
38.7% if respondents perceived minor bias, to 52.3% and
to an even greater extent, when respondents perceived
major bias (78.4%) (Table 6).
The statistical significance of a relationship between bias
perception and regulatory preferences (p-value 0.002
–
0.012) is established. The findings underscore mounting
discontent around the issue of AI fairness and growing
need for regulators to step in especially to those who
discern discriminatory patterns in AI in decision making.
Table 6: AI Underwriting Bias Perception vs. Regulation Preferences
Bias Perception
Favor Strong AI
Regulations (%)
Favor Minimal AI
Regulations (%)
p-value
No bias
21.5%
45.3%
0.012
Minor bias
52.3%
30.2%
0.008
Major bias
78.4%
10.5%
0.002
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
17
Figure 6: Impact of Bias Perception on AI Regulation Preferences
Economic Benefits of AI-Powered Underwriting
The study also looked into whether the AI driven
underwriting is efficient and economically beneficial.
Previous to AI implementation, claim processing time
averaged 10 days and with AI implementation the average
was reduced to 3 days. It shows the AI model’s operational
efficiency resulting in this significant 70% (p = 0.002)
reduction of the time taken to process claims.
The underwriting cost per policy decreased from $500 to
$200 (p = 0.004), a significant amount of underwriting cost
reduction. Increasing sales amount (from 0 to 20),
decreased reporting time (3 weeks to 2 weeks) and
increased underwriting accuracy from 82% to 92% (p =
0.001)
(Table 7) supported AI’s accuracy enhancing ability
with regard to actuarial models.
Insurers can significantly benefit from the AI underwriting
innovation as these results indicate that AI underwriting
helps insurers reduce costs, process faster and improve
accuracy.
Table 7: Economic Benefits of AI-Powered Underwriting
Economic Indicator
Before AI
Implementation
After AI
Implementation
p-value
Reduction in Claim
Processing Time
10 days
3 days
0.002
Cost Savings per
Policy ($)
$500
$200
0.004
Increase in
Underwriting
Accuracy (%)
82%
92%
0.001
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
18
Figure 7: Economic Impact of AI Implementation
Consumer Trust in AI vs. Human Underwriting
Although operational and economic benefits of AI show up
already, consumer trust is a big hurdle to complete
adoption of AI in insurance underwriting. The consumer
trust level for AI powered underwriting model was 3.2 on
5 and only 28.4% of participant preferred AI underwriting
models (Table 8).
The trust score for human based underwriting was of 4.1
with 45.2% of the participants preferring to have
traditional human underwritten policies. Hybrid AI-
human underwriting registered trust levels highest among
the other approaches, i.e, 4.5 on the trust scale and while
62.1% of the participants favored a blended AI-human
approach. The significant differences (p-value of 0.003
–
0.018) in consumer preferences confirm the skepticism of
consumers to fully autonomous AI underwriting.
The results of this article indicate that adopting these
hybrid models will allow organizations facing trust issues
and wanting to use AI as a technology, to obtain both AI’s
analytical power and human expertise.
Table 8: Consumer Trust in AI vs. Human Underwriting
Underwriting
Type
Consumer Trust
Level (1-5)
Percentage
Preferring This
Approach (%)
p-value
AI-Powered
3.2
28.4%
0.018
Human-Based
4.1
45.2%
0.007
Hybrid (AI +
Human)
4.5
62.1%
0.003
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
19
Figure 8: Consumer Trust and Preference in Underwriting Approaches
AI Bias and Its Impact on Perceived Efficiency
The study utilized a chi-square test to examine the effect
of perceived AI bias to efficiency perceptions. As shown in
Table 9, respondents who perceived AI underwriting as
having no bias had a rating of 4.2 out of 5 while those from
whom it was perceived to be set over a minor bias gave a
rating of 3.5. In the case of participants who found the
major bias, the efficiency rating fell to 2.8 and to an
efficiency rating of 2.1 for the perception of extreme bias.
Respondents who expressed racial bias concern rated AI
efficiency as 3.0 while those concerned with gender rated
it as 3.3. Using chi-
square test results (χ² = 15.67, p =
0.0003) it becomes evident that there exists a strong
statistical tie between AI bias perception and efficiency
ratings. In essence, these findings suggest that bias
concerns have a profound negative effect on attitude
towards AI effectiveness and highlight the necessity of
incorporating bias mitigation strategies into underwriting
models (Table 9).
Table 9: Chi-Square Test - AI Bias vs. AI Efficiency Perception
AI Bias Perception
Avg AI Efficiency
Score (1-5)
Chi-Square
Statistic
p-value
No bias
4.2
6.43
0.011
Minor bias
3.5
8.91
0.004
Major bias
2.8
12.35
0.001
Extreme bias
2.1
15.67
0.0003
Racial bias concerns
3.0
9.21
0.006
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
20
Gender bias
concerns
3.3
7.88
0.008
Figure 9: Impact of AI Bias Perception on Efficiency Scores
AI Familiarity and Trust in Underwriting Decisions
An ANOVA test is conducted to find out whether there is a
relationship between familiarity with AI and the trust of AI
in underwriting. These results demonstrate that people’s
trust in AI models increases with increasing familiarity
with the models. People not previously familiar with AI
underwriting scored their trust in 2.5/5 while marginally
familiar scored 3.8/5. Trust levels among very familiar
respondents sit at 4.2 and where respondents are AI
experts, the rating hits 4.6.
Frequent users of AI were given a score of 4.0 and trust
remained the highest at 4.8 in cases of AI research
professionals, which is unsurprising. In order to determine
the existence of a statistically significant difference in trust
among different levels of familiarity with AI, the F-statistic
(F = 13.27, p = 0.0002) clearly shows that such a difference
in trust exists among various familiarity levels (Table 10).
Table 10: ANOVA - AI Familiarity vs. Trust in AI Underwriting
AI Familiarity
Level
Avg Trust in AI
Underwriting (1-5)
F-Statistic
p-value
Not familiar
2.5
5.32
0.009
Somewhat familiar
3.8
8.21
0.002
Very familiar
4.2
10.45
0.001
Expert
4.6
12.89
0.0006
Frequent AI user
4.0
11.54
0.0008
AI research
professional
4.8
13.27
0.0002
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
21
Figure 10: Trust in AI Underwriting Based on AI Familiarity
The Impact of AI Efficiency on Market Growth and Cost
Savings
The impact of AI efficiency on key financial and
operational metrics in underwriting was measured using
a regression analysis. The analysis in Table 11 shows that
reductions in AI efficiency score is associated with
improvements to cost per policy through regression, with
the coef
ficient (β) =
-120.5 (p = 0.001), where greater AI
efficiency results in lower cost per policy.
Lastly, the ability for market growth to increase with the
complexity of the AI model (β = 15.8, p = 0.002) is
indicative of development within the industry towards
more advanced AI driven actuarial models.
Similarly, AI training data quality was also a significant
factor in predicting fraud detection accuracy (β = 8.3, p =
0.0005) which reinforces the fact that the better the
quality of the data used for training is, the lesser
underwriting risks for the lender. Deepening this point in
the context of claims processing, AI automation had a
pronounced effect on underwriting speed (β = 22.1, p =
0.0008) in line with the notion of automation’s positive
impact on operational efficiency.
Interestingly, AI data privacy strength was also related to
consumer trust in the same direction (β = 5.9, p = 0.003).
These findings shed light on the economic, operational and
consumer trust benefits of the AI underwriting and stress
on the need for ensuring data quality, complexity of the
model and regulatory compliance for the full effectiveness
of AI (Table 11).
Table 11: Regression Analysis - AI Efficiency & Market Growth
Independent
Variable
Dependent
Variable
Regression
Coefficient
(β)
Standard
Error
R-Squared
p-value
AI Efficiency
Score
Cost Savings
per Policy ($)
-120.5
15.2
0.82
0.001
AI Model
Complexity
Market
Growth (%)
15.8
3.7
0.76
0.002
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
22
AI Training
Data Quality
Fraud
Detection
Accuracy (%)
8.3
2.9
0.85
0.0005
AI
Automation
in Claims
Underwriting
Speed
Increase (%)
22.1
4.2
0.79
0.0008
AI Data
Privacy
Strength
Consumer
Trust Score
5.9
1.7
0.68
0.003
Figure 11: Regression Analysis of AI Factors on Key Metrics
Consumer Trust in AI vs. Human Underwriting
To compare the consumer trust in AI Underwriting versus
Human Underwriting, hybrid models and other alternative
models, a t- test analysis was performed. As indicated in
the results in Table 12, human underwriting was
significantly preferred compared to AI-only underwriting
(mean trust = 4.1, p = 0.0004).
Hybrid AI and human underwriting models were bitwise
trusted most (4.5), which was statistically significantly
different compared to the trust scores associated with
both AI only (3.3, p < 0.002) and human only (3.5, p <
0.002) approaches. Regarding the difference between
regulated AI vs. unregulated AI, the results were that
regulated AI (mean trust = 4.3) was trusted more than
unregulated AI (mean trust = 2.9, p = 0.0002).
There was also a greater mean trust placed in the
blockchain based risk assessment (mean trust = 4.2)
compared to that of AI only underwriting's (mean trust =
3.6, p = 0.0050). Findings show that consumers are
skeptical about standalone AI underwriting but are more
accepting with models that have some form of human
oversight, regulation or one of the decentralized
verifications such as blockchain (Table 12).
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
23
Table 12: T-Test - Consumer Trust in AI vs. Human Underwriting
Comparison
Mean Trust in
AI (1-5)
Mean Trust in
Other (1-5)
T-Statistic
p-value
AI
Underwriting
vs. Human
Underwriting
3.2
4.1
5.89
0.0004
AI
Underwriting
vs. Hybrid (AI
+ Human)
3.2
4.5
4.72
0.0020
Human
Underwriting
vs. Hybrid (AI
+ Human)
4.1
4.5
2.85
0.0140
AI
Underwriting
(Unregulated)
vs. AI
(Regulated)
2.9
4.3
6.23
0.0002
AI vs.
Blockchain-
based Risk
Assessment
3.6
4.2
3.78
0.0050
Figure 12: Trust in AI vs. Other Methods
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
24
Factors Influencing Consumer
Trust
in
AI
Underwriting
The key factors of consumer trust in AI underwriting
models were explored using a logistic regression model. As
shown in Table 13, AI transparency (Odds Ratio = 2.3, p =
0.002) and AI fairness (Odds Ratio = 1.8, p = 0.004) were
both significant prognosticants for consumer trust,
highlighting the need that consumers place on
transparency and ethical practice of AI.
Trust was positively associated with AI efficiency (Odds
Ratio = 2.5, p = 0.001) and users trusted AI to the extent
that it improves underwriting outcomes. The most
significant factor determined the impact of trust on the
regulators (Odds Ratio = 3.1, p = 0.0005), which means
that strict AI governance is significantly important for
consumer acceptance.
Even consumer AI knowledge (Odds Ratio = 1.6, p = 0.007)
was a significant factor, signifying that consumers who are
better informed are more likely to trust AI driven
underwriting (Table 13).
Table 13: Logistic Regression - Predicting AI Trust Based on Key Factors
Independent
Variable
Odds Ratio
Standard Error
95%
Confidence
Interval
p-value
AI
Transparency
2.3
0.4
(1.7, 2.9)
0.002
AI Fairness
1.8
0.3
(1.4, 2.2)
0.004
AI Efficiency
2.5
0.5
(2.0, 3.0)
0.001
Regulatory
Oversight
3.1
0.6
(2.5, 3.7)
0.0005
Consumer AI
Knowledge
1.6
0.3
(1.3, 1.9)
0.007
Figure 13: AI Factors and Their Odds Ratios
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
25
Performance Comparisons: AI-Driven vs. Traditional
Underwriting Approaches
Various underwriting modelling tools namely AI driven,
traditional, hybrid and also alternative approach were
compared through a descriptive statistical analysis.
From Table 14, it is noticed that fully automated AI
underwriting completed in minimum processing time (2.0
days) with highest fraud detection rate (96.5%). While the
fastest methods, AI-human underwriting (3.1 days, 93.0%
fraud) and AI (none, 92.8% fraud), exhibited the lowest
effectiveness and regulatory compliance, along with the
least confidence for consumers, it remains that speed is
not the determinant of effectiveness.
Blockchain based underwriting models had one of the
highest fraud detection rates (95.0%) largely due to the
improvements in data security and verification process.
Fully automated AI underwriting is most efficient, hybrid
models offer the right balance between accuracy,
compliance and customer trust and blockchain can be used
in underwriting to enhance the performance upon
accuracy and speed (on the second delivery) while
maintaining the same compliance (Table 14).
Table 14: Descriptive Statistics - AI-Driven vs. Traditional Underwriting Outcomes
Underwriting
Approach
Avg Processing
Time (Days)
Avg Fraud
Detection Rate (%)
p-value
AI-Driven
2.5
92.5
0.001
Traditional
10.8
85.3
0.003
Hybrid (AI +
Human)
5.2
89.7
0.0008
AI with
Explainability
Features
3.1
93.0
0.0005
AI with Regulatory
Compliance
4.0
91.5
0.002
AI with Deep
Learning Models
2.8
94.2
0.0009
AI with Limited
Human Oversight
3.5
90.3
0.0015
AI with Consumer
Feedback
Integration
3.3
92.1
0.0007
Blockchain-Based
Underwriting
4.5
95.0
0.0012
Fully Automated
AI Underwriting
2.0
96.5
0.0004
DISCUSSION
The Role of AI in Insurance Underwriting: Balancing
Efficiency and Fairness
Integration of AI powered actuarial models into U.S.
insurance underwriting has resulted in great
improvement in the assessment of risk, fraud detection
and operation (Mishra, 2024; Paul, 2024). Bearing this in
mind and according to the existing literature (Pugnetti &
Seitz, 2021; Singh & Gautam, 2024), the mentioned use of
AI in underwriting shortens the processing time and
increases fraud discovery. These benefits have seen
consumers’ skepticism, regulatory scr
utiny and ethical
issues, especially regarding bias and transparency (2024
du Preez et al, 2025 Umar & Reuben).
In health, life, property and casualty insurance sectors in
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
26
the U.S, AI is being adopted and insurers use big data
analytics and machine learning to fine grained
segmentation of their risks and accuracy in pricing
(Anbalagan, 2024; Srirangam et al, 2024). This study
confirms the results that hybrid (AI
–
human) and regulated
AI underwriting approaches are considered to be more
trusted, especially compared to purely AI based
underwriting approaches and that they are considered
more efficient. These findings point out the existing
tension between efficiency and fairness as one of the core
issues to address in the adoption of AI in U.S. actuaries' and
regulators' frameworks (Kharlamova et al, 2024).
AI Efficiency vs. Trust: A Persistent Trade-Off
The trade off with respect to the efficiency of using AI and
consumer distrust is one of the most striking results of this
study. In the U.S. market, AI underwriting models have
proved to shorten claim processing time from 10 days to
even 2-3 days (Table 7) but consumers still are not
comfortable with full automated AI underwriting (Table
12). Aparis (2024) suggests that it is a common knowledge
that automation can speed-up things, professional
judgment needed to build trust is missing there (62.1%
preference for hybrid AI-human models over 28.4% for AI
only models compared) as shown in Kumar (2024).
The structure of trust is shaped by the rules of the game,
i.e. regulation. In Table 12 it can be seen that more trust is
expressed in regulated AI underwriting (mean = 4.3) than
in unregulated AI models (mean = 2.9, p = 0.0002). This
mirrors existing research in showing the way to
algorithmic auditing, regulatory oversight and fairness
and reduction of bias in automated decision making
(O’Neil et al, 2024, Chandler, 2025). Regulation is slowly
starting to play a part in this, including by agencies such as
Consumer Financial Protection Bureau (CFPB) and the
National Association of Insurance Commissioners (NAIC),
who have been urging for explainable AI and mitigating
the potential sources of bias in the process of designing
machines (Tumai, 2021; Pareek, 2023).
Results also support that the trust from consumer is based
on
the consumer’s familiarity with the AI (Table 10).
Compared to frequent AI users and AI research
professionals, far fewer indicated higher levels of trust
(mean = 4.8, p = 0.0002). Consumer skepticism can be
reduced and AI adoption might increase if AI literacy is
increased by strengthening transparency initiatives
among the consumers (Singh & Gautam, 2024; Umar &
Reuben, 2025).
Bias and Fairness Concerns in AI-Driven Underwriting
While true that AI has the potential to lower the human
element that brings subjectivity to decision making,
algorithmic bias in U.S. insurance underwriting remains a
concern. Table 9 using this study supports that perceived
bias has a negative effect on ratings of AI efficiency and the
lowest efficiency scores (2.1 out of 5, p = 0.0003) are for
extreme cases of bias. In the U.S. life and health insurance
sectors, AI bias concerns have been widely documented
with racial and gender bias in risk assessment models
widely established (Adeniran et al, 2024; Pareek, 2023).
As shown in the Chi-Square results (Table 9), those who
believe AI underwriting is racially and/or gender biased
have significantly smaller efficiency scores (p = 0.01). In
line with previous studies showing that biased training
data poses risks for credit and insurance score and AI
models are also not explainable (O’Neil et al, 2024; Zarifis
& Cheng, 2021), this result makes sense. The results also
correspond with the action taken by the U.S. regulatory
agencies like the New York Department of Financial
Services (NYDFS) that directed companies offering
insurance products in the state to implement bias auditing
and fairness testing to ensure their AI systems are not
being biased during customers’ underwriting processes
(Chandler, 2025, Pareek, 2023).
An option to this is using explainable AI (XAI) models so
insurers and regulators can audit the AI driven decisions
and reduce discriminatory outcomes (O’Neil et al, 2024)
(Umar & Reuben, 2025). Achieving this comes with higher
investment in algorithmic transparency, something that
many US insurers still struggle with because of their many
proprietary black box models (King et al, 2021)
The Economic Trade-Offs: Cost Savings vs. Market
Adoption
The economic implications of AI applied to risk
assessment are of consequence. Results confirm that AI
read writing significantly reduces cost (p = 0.004) (Table
7) to $200 vs. $500 per policy. Industry reports indicate
that insurers utilizing predictive analytics have 20-30%
cost of operational savings (Mishra, 2024; Anbalagan,
2024) and this aligns in line with the same.
Market adoption isn’t a challenge while it is evident that
cost efficiency is there. Table 6 shows the stronger
people’s perception of AI bias, the stronger their support
for AI regulators (p = 0.002), that is, 78.4% of the
respondents who perceived major AI bias supported
strong AI regulators. Without having adequate fairness
and accountability frameworks in place, insurers face
regulatory push back and decreased consumer adoption
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
27
(Chandler, 2025; Kumar, 2024).
As presented in Table 14, fully automated AI underwriting
proves to be the fastest (2.0 days) and most accurate
(96.5% fraud detection) compared with the hybrid and
blockchain models, they demonstrate an advantage in
terms of compliance with ethical and regulatory
standards. The underwriting based on blockchain also
means the underwriting will grow (fraud detection =
95.0%, p = 0.0012) and the future risk assessment may
depend on the decentralized, tamper-proof data
verification to reduce some biases (Vandervorst et al,
2022; Taneja et al, 2024).
How AI-Powered Underwriting Benefits the U.S.
Economy, Health, Security and Technology
The broad implications of the findings of this study are for
the U.S. economy, public health, security and technological
advancement. The capability of AI for risk assessment in
insurance underwriting can achieve this purpose by
optimizing financial efficiency, preventing fraud, enable
healthcare accessibility and inform public policy decision.
The utilization of AI for actuarial models provides a more
efficient, less expensive and a more competitive U.S.
insurance industry by reducing underwriting costs by
60+% (Table 7) and claim processing time from 10 to 3
days (MUPA et al, 2025). This is in line with the rising trend
of automation in financial risk management, as insurers
with machine learning models enjoy substantial cuts to
their underwriting overhead (Yadav & Bank).
From a healthcare standpoint, the use of AI in
underwriting helps to better risk stratify a person’s risk
and provide a more tailored and affordable insurance
policy to individuals who would otherwise not be able to
afford it, especially high-risk individuals (Oberkrome,
2023). Medical insurance underwriting with predictive
analytics helps the medical insurers to structure a better
policy that fits patients with chronic conditions, hence
reducing the rate of uninsurance in the U.S and even
medical insurance bankruptcy cases. (Patil et al, 2023).
Utilization of AI for health insurance fraud detection saves
billions in fraudulent claims, contributing to the right
funding to actual beneficiaries (Larzelere, 2021).
AI based underwriting helps in detection of financial
crimes, from national security and fraud prevention
perspectives, by detecting of data misrepresentation
patterns in insurance applications (Patil et al, 2023).
Given, the U.S. economy loses over $308 billion dollars
annually on insurance fraud, enhanced by the AI powered
models in real time fraudulent claims detection and claims
verification (MUPA et al, 2025).
With regard to technological and commercialization
aspects, AI in underwriting runs true to the fancy of the
Insurtech industry that is expected to hit $20 billion in
2028 (Yadav & Bank). Blockchain based risk assessment
(Table 14) has tamper proof underwriting records which
fall in line with regulatory requirement and dispense the
possibility of disputes around AI (Oberkrome, 2023). This
technology has commercial potential in both traditional
insurance markets as well as newer markets including
cybersecurity insurance, climate risk assessment and gig
economy coverage (Patil et al, 2023).
The results of this study point out that public policy actions
must be taken to guarantee AI fairness and transparency
and accountability in underwriting. Policymakers need to
determine the regulatory frameworks that ought to be in
place to enhance efficiency in the use of AI while protecting
consumers from discriminatory outcomes brought about
by automated models that could unfairly marginalize
already disadvantaged populations (Larzelere, 2021).
Following the discussions of the legislation in Congress
and the regulatory agencies (Federal Trade Commission
(FTC) and the National Association of Insurance
Commissioners
(NAIC)),
recently,
AI
auditing
requirements become a trend to encourage the fairness in
the insurance risk assessment (MUPA et al, 2025).
AI in underwriting contributes towards economic growth
by reducing inefficiencies, also towards healthcare by
making policy affordable, strengthens financial security
through fraud d
etection and to USA’s technological
leadership (or at least has potential) in the Insurtech
sector. AI underwriting is likely to gain broad public trust
and acceptance only if it is duly regulated, fairly applied
and accompanied by educational initiatives for consumers
(Patil et al, 2023; Yadav & Bank).
Future Research and Policy Implications
The findings of this study underscore several important
policy implications for the U.S. insurance industry:
1. Regulatory Auditing
–
U.S. regulators should
implement mandatory AI bias audits and explainability
standards to ensure fairness in underwriting decisions
(O’Neil et al, 2024; Chandler, 2025).
2. Hybrid AI-Human Models
–
To balance efficiency
and trust, insurers should adopt AI-human collaboration
frameworks for underwriting (Apergis, 2024; Umar &
Reuben, 2025).
3. Consumer AI Literacy Initiatives
–
Educating
consumers on AI models, risk assessment methods and
bias detection could increase trust and adoption (Singh &
Gautam, 2024; Kumar, 2024).
4. Blockchain for Risk Assessment
–
Blockchain
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
28
based underwriting models help in securing data, the
prevention of frauds and reduction of bias (Taneja et al,
2024; Vandervorst et al, 2022).
Comparison between these insights and their impact on
the current academic and industry debate on AI-driven
underwriting suggests that responsible AI development is
a necessary prerequisite for full utilization of the potential
of AI in the U.S. insurance market with respect to both
efficiency and fairness.
CONCLUSION
The results of this study show that for U.S insurance
underwriting, AI powered actuarial models can hold a
great deal of transformative power in increasing
efficiency, fraud detection and decreasing costs. AI driven
underwriting has been shown to ease the process of
making the decision, increasing the pace of claim
processing and cut underwriting expenses, as well as
improve the accuracy of fraud detection. These advances
make the insurance industry more efficient and
competitive on the strength of the insure
rs’ ability to
determine risks with greater accuracy.
The study identifies trust, transparency and fairness as
long-suffering areas. Skepticism from consumers on the
part of AI models act as a key barrier in the widespread
adoption. Hybrid AI human underwriting models is
preferred for the reason that combining human oversight
in AI underwriting can lead to greater trust, clarifying
concerns on the fairness and reliability of the automated
underwriting decisions. The regulatory oversight has
proven to be a critical factor influencing consumer
confidence in Artificial Intelligence underwriting which
concludes that the public trust in AI underwriting is
significant when regulatory frameworks will assure
transparency and fairness.
The issue of algorithmic bias is still important and
respondents who see racial and gender bias in AI
underwriting models give significantly lower efficiency
ratings. Indeed, these findings are comparable to
prevailing apprehensions with respect to biased training
data and unknowable decision making in AI applications.
To solve these challenges, algorithmic fairness for AI has
to be committed, explaining AI models and regulatory
frameworks that hold underwriting decisions accountable.
A business case using AI in underwriting is clear on
economic purposes
–
cost saving, fraud reduction and
market expansion. The full potential of AI in insurance
underwriting can be realized only when the tradeoff
between the efficiency of the AI application and the trust
of the consumer and regulatory compliance is made. AI
and blockchain technology have been emerging as a
promising future that can complement each other in
creating a safer data security, fraud prevention and
verifiability in underwriting decisions. Integrating
blockchain based underwriting models can be another
way for AI driven insurance policy to have greater
transparency and fairness in the assessment of risk; which
in a way may add credibility to AI insurance policies.
This study points out some key recommendations that
help realize the benefits fully from AI in insurance
underwriting. Important will be the development and
enforcement of regulatory policies to mitigate bias, for
algorithmic auditing and explainability in order to ensure
ethical AI adoption. While the efficiency benefits of AI are
important, insurers may benefit from keeping human
judgement in complex risk assessment cases; hybrid AI-
human underwriting models could help to ensure that the
case is handled efficiently while erring on the side of
caution. In improving public confidence in AI driven
underwriting, AI literacy needs to be improved through
consumer education initiatives.
This study contributes the broader discussion on the
future of AI in the insurance underwriting, having proven
that although AI opens the window of efficiency and
innovation, its wide adoption should be based on the
principles of fairness, transparency and accountability.
The insurer’s opportunity to responsibly deploy AI serves
to create a more inclusive and efficient and consumer
centric underwriting landscape within the U.S. insurance
industry. Resolution of the challenges proffered in this
study, will drive economic growth, foster trust and place
the U.S. in the front position as a leader in future insurance
technology.
REFERENCES
Abuamoud, I., Lillywhite, J., Simonsen, J., & Al-Oun, M.
(2016). Factors influencing food security in less popular
tourists sites in Jordan's Northern Badia. International
Review of Social Sciences and Humanities, 11(2), 20-36.
Adeniran, I. A, Efunniyi, C. P, Osundare, O. S, Abhulimen, A.
O, & OneAdvanced, U. K. (2024). Advancements in
predictive modeling for insurance pricing: Enhancing risk
assessment and customer segmentation. International
Journal of Management & Entrepreneurship Research,
6(8).
Afshar, M. Z. (2023). Exploring Factors Impacting
Organizational Adaptation Capacity of Punjab Agriculture
& Meat Company (PAMCO). International Journal of
Emerging Issues in Social Science, Arts and Humanities
(IJEISSAH),
2(1),
1-10.
https://doi.org/10.60072/ijeissah.2023.v2i01.001
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
29
Ahmad, S. (2025). Entrepreneurship and Sustainable
Leadership Practices: Examine how entrepreneurial
leaders incorporate sustainability into their business
models and the leadership traits facilitating this
integration. Journal of Entrepreneurship and Business
Venturing, 5(1).
Ahmed, A., Rahman, S., Islam, M., Chowdhury, F., & Badhan,
I. A. (2023). Challenges and Opportunities in
Implementing Machine Learning For Healthcare Supply
Chain Optimization: A Data-Driven Examination.
International journal of business and management
sciences, 3(07), 6-31.
Anbalagan, K. (2024). Cloud-Powered Predictive Analytics
in Insurance: Advancing Risk Assessment Through AI
Integration. INTERNATIONAL JOURNAL OF ENGINEERING
AND TECHNOLOGY RESEARCH (IJETR), 9(2), 195-206.
Apergis, A. (2024). The role of cognitive assistants in
medical insurance underwriting.
Badhan, I. A., Hasnain, M. N., Rahman, M. H., Chowdhury, I.,
& Sayem, M. A. (2024). Strategic Deployment of Advance
Surveillance Ecosystems: An Analytical Study on
Mitigating Unauthorized US Border Entry. Inverge Journal
of Social Sciences, 3(4), 82-94.
Badhan, I. A., Neeroj, M. H., & Rahman, S. (2024). Currency
rate fluctuations and their impact on supply chain risk
management: An empirical analysis. International journal
of business and management sciences, 4(10), 6-26.
Butt, S., & Yazdani, N. (2023). Relationship Between
Execution of Quality Management Practices and Firm’s
Innovation Performance: A Review of Literature. Journal
of Asian Development Studies, 12(3), 432-451.
Butt, S., Umair, T., & Tajammal, R. (2024). Nexus between
Key Determinants of Service Quality and Students’
Satisfaction in Higher Education Institutions (HEIs).
Annals of Human and Social Sciences, 5(2), 659-671.
Chandler, S. J. (2025). Compared to What? The Paramount
Question in the Regulation of Medical AI. Houston Journal
of Health Law & Policy, 24(1), 1-73.
Dhal, K., Karmokar, P., Chakravarthy, A. et al. Vision-Based
Guidance for Tracking Multiple Dynamic Objects. J Intell
Robot
Syst
105,
66
(2022).
https://doi.org/10.1007/s10846-022-01657-6
Dixit, S., & Jangid, J. (2024). Exploring Smart Contracts and
Artificial Intelligence in FinTech. https://jisem-
journal.com/index.php/journal/article/view/2208
du Preez, V, Bennet, S, Byrne, M, Couloumy, A, Das, A,
Dessain, J, ... & van Heerden, L. (2024). From bias to black
boxes: understanding and managing the risks of AI
–
an
actuarial perspective. British Actuarial Journal, 29, e6.
Easwaran, V., Orayj, K., Goruntla, N., Mekala, J. S.,
Bommireddy, B. R., Mopuri, B., ... & Bandaru, V. (2025).
Depression, anxiety, and stress among HIV-positive
pregnant women during the COVID-19 pandemic: a
hospital-based cross-sectional study in India. BMC
Pregnancy and Childbirth, 25(1), 134.
Hood, K., & Al-Oun, M. (2014). Changing performance
traditions and Bedouin identity in the North Badiya,
Jordan. Nomadic Peoples, 18(2), 78-99.
Jagdish Jangid. (2023). Enhancing Security and Efficiency
in Wireless Mobile Networks through Blockchain.
International Journal of Intelligent Systems and
Applications in Engineering, 11(4), 958
–
969. Retrieved
from
https://ijisae.org/index.php/IJISAE/article/view/7309
Kaniz, R. E., Lindon, A. R., Rahman, M. A., Hasan, M. A., &
Hossain, A. (2025). THE IMPACT OF PROJECT
MANAGEMENT STRATEGIES ON THE EFFECTIVENESS OF
DIGITAL MARKETING ANALYTICS FOR START-UP
GROWTH IN THE UNITED STATES. project management,
4(1).
Kharlamova, A, Kruglov, A, & Succi, G. (2024, May). State-
of-the-Art Review of Life Insurtech: Machine learning for
underwriting decisions and a Shift Toward Data-Driven,
Society-oriented Environment. In 2024 International
Congress on Human-Computer Interaction, Optimization
and Robotic Applications (HORA) (pp. 1-12). IEEE.
King, M. R, Timms, P. D, & Rubin, T. H. (2021). Use of big
data in insurance. The Palgrave Handbook of
Technological Finance, 669-700.
Kumar, S. (2024). Navigating the Complexities of
Insurance Underwriting Results through Artificial
Intelligence. International Journal of Science and Research
(IJSR), 13(4), 1464-1471.
Larzelere, S. P. (2021). The Cognitive and Emotional
Reactions of Commercial Casualty Insurance Underwriters
to the Use of Predictive Analytics (Doctoral dissertation,
University of Pennsylvania).
Mishra, B. (2024). Machine Learning for Financial
Professionals. Educohack Press.
Muhammad Zaurez Afshar1*, Dr. Mutahir Hussain Shah2.
(2025). Strategic Evaluation Using PESTLE and SWOT
Frameworks: Public Sector Perspective. ISRG Journal of
Economics, Business & Management (ISRGJEBM), III(I),
108
–
114. https://doi.org/10.5281/zenodo.14854362
MUPA, M. N, TAFIRENYIKA, S, RUDAVIRO, M, NYAJEKA, T,
MOYO, M, & ZHUWANKINYU, E. K. (2025). Machine
Learning in Actuarial Science: Enhancing Predictive
Models for Insurance Risk Management.
Nguyen, H. U., Trinh, T. X., Duong, K. H., & Tran, V. H.
(2018). Effectiveness of green muscardine fungus
Metarhizium anisopliae and some insecticides on lesser
coconut
weevil
Diocalandra
frumenti
Fabricius
(Coleoptera: Curculionidae). CTU Journal of Innovation
and Sustainable Development, (10), 1-7.
Nguyen, L., Trinh, X. T., Trinh, H., Tran, D. H., & Nguyen, C.
(2018). BWTaligner: a genome short-read aligner.
Vietnam Journal of Science, Technology and Engineering,
60(2), 73-77.
Frontline Marketing, Management and Economics Journal
FRONTLINE JOURNALS
30
Oberkrome, F. (2023). Predictive Analytics in Health
Insurance: Enhancing Risk Assessment and Policy
Customization.
O'Neil, C, Sargeant, H, & Appel, J. (2024). Explainable
fairness in regulatory algorithmic auditing. W. Va. L. Rev,
127, 79.
Pareek, C. S, & Heights, B. Enhancing Quality Assurance in
Annuities: A Risk Management Approach with AI and
Machine Learning.
Pareek, C. S. Unmasking Bias: A Framework for Testing and
Mitigating AI Bias in Insurance Underwriting Models. J
Artif Intell. Mach Learn & Data Sci 2023, 1(1), 1736-1741.
Patil, S, Patil, A, Patil, V, Yadav, S, & Das, S. (2023,
December).
Cognitive
Data
Underwriting
with
Automation: A Survey. In International Conference on
Business Data Analytics (pp. 287-298). Cham: Springer
Nature Switzerland.
Paul, J. (2024). AI-Powered Data Analytics: Shaping the
Future of Auto Insurance Pricing and Claims Processing.
Pugnetti, C, & Seitz, M. (2021). Data-driven services in
insurance: Potential evolution and impact in the Swiss
market. Journal of Risk and Financial Management, 14(5),
227.
Rahman, S., Sayem, A., Alve, S. E., Islam, M. S., Islam, M. M.,
Ahmed, A., & Kamruzzaman, M. (2024). The role of AI, big
data and predictive analytics in mitigating unemployment
insurance fraud. International Journal of Business
Ecosystem & Strategy (2687-2293), 6(4), 253-270.
Rasul, I., Akter, T., Akter, S., Eshra, S. A., & Hossain, A.
(2025). AI-Driven Business Analytics for Product
Development: A Survey of Techniques and Outcomes in
the Tech Industry. Frontline Marketing, Management and
Economics Journal, 5(01), 16-38.
Sachin Dixit, & Jagdish Jangid. (2024). Asynchronous SCIM
Profile for Security Event Tokens. Journal of
Computational Analysis and Applications (JoCAAA),
33(06),
1357
–
1371.
Retrieved
from
https://eudoxuspress.com/index.php/pub/article/view/
1935
Sawyer, S., Ellers, S., Kakumanu, M. S., Bommireddy, B.,
Pasgar, M., Susan-Kurian, D., ... & Jurdi, A. A. (2025). Trial
in progress for a colorectal cancer detection blood test.
https://ascopubs.org/doi/10.1200/JCO.2025.43.4_suppl.
TPS306
Shabbir, A., Arshad, N., Rahman, S., Sayem, M. A., &
Chowdhury, F. (2024). Analyzing surveillance videos in
real-time using AI-powered deep learning techniques.
International Journal on Recent and Innovation Trends in
Computing and Communication, 12(2), 950-960.
Singh, D, & Gautam, A. (2024). Unlocking the Power of Big
Data in Insurance: The Role of Data Analytics. In Data
Alchemy in the Insurance Industry: The Transformative
Power of Big Data Analytics (pp. 13-26). Emerald
Publishing Limited.
Srirangam, R. K, Chennuri, S, & Pendyala, V. (2024).
Technological disruption in P&C insurance: The impact of
advanced analytics on risk assessment and customer
engagement. International Journal of Research in
Computer Applications and Information Technology
(IJRCAIT), 7(2), 1224-1237.
Taneja, S, Bisht, V, & Kukreti, M. (2024). Revolutionizing
Insurance Practices Through Advanced Data Alchemy. In
Data Alchemy in the Insurance Industry: The
Transformative Power of Big Data Analytics (pp. 119-131).
Emerald Publishing Limited.
Tumai, W. J. (2021). A critical examination of the legal
implications of Artificial Intelligence (AI) based
technologies in New Zealand workplaces (Doctoral
dissertation, The University of Waikato).
Umar, J, & Reuben, J. (2025). Fair AI in Real Estate and
Insurance: Overcoming Bias and Improving Transparency
in Machine Learning Models.
Vandervorst, F, Verbeke, W, & Verdonck, T. (2022). Data
misrepresentation detection for insurance underwriting
fraud prevention. Decision Support Systems, 159, 113798.
Yadav, S, & Bank, S. V. OPTIMIZING LIFE INSURANCE RISK
ASSESSMENT THROUGH MACHINE LEARNING A DATA-
DRIVEN APPROACH TO PREDICTIVE UNDERWRITING.
Zarifis, A, & Cheng, X. (2021, July). Evaluating the new AI
and data driven insurance business models for
incumbents and disruptors: Is there convergence?. In
Business Information Systems (pp. 199-208).
Zaurez Afshar, M., & Hussain Shah, M. (2025). Performance
Evaluation Using Balanced Scorecard Framework: Insights
from A Public Sector Case Study. INTERNATIONAL
JOURNAL OF HUMAN AND SOCIETY, 5(01), 40
–
47.
https://ijhs.com.pk/index.php/IJHS/article/view/808
