Page 70 - IMDR Journal 2025

Research Article
practices for improving transparency and trust in AI-driven 1. Convenience Sampling Bias
systems. The use of convenience sampling may introduce bias, as
Sampling Framework respondents may not fully represent the broader population
Sample Frame The study targets health insurance of policyholders. For example, urban respondents are
policyholders in Pune, India, who have recently filed claims. overrepresented in the sample, which may limit the
Pune is chosen as the study location due to its mix of urban generalizability of findings to rural areas.
and semi-urban populations, making it representative of 2. Self-Reported Biases
India's diverse insurance market. Respondents are selected Survey responses are based on self-reported data, which
based on their experience with claims processing, ensuring may be subject to biases such as social desirability bias or
that they have first-hand insights into the efficiency, fairness,
recall bias. Respondents may overstate their satisfaction or
and transparency of AI-driven systems.
underreport challenges due to perceived expectations.
Sample Size A sample size of approximately 150 3. Limited Scope
respondents is selected to ensure a balance between depth of
analysis and feasibility. This size is sufficient to identify The study focuses on health insurance policyholders in
trends and patterns in consumer perceptions while Pune, which may not fully capture the experiences of
remaining manageable within the study's scope. The sample policyholders in other regions or with different types of
size is determined based on the need for statistical insurance.
significance and the availability of respondents. 4. Technical Limitations
Sampling Technique The study uses convenience The study does not evaluate the technical aspects of AI
sampling, a non-probability sampling technique, to recruit algorithms or their development, limiting its ability to
respondents. Convenience sampling is chosen due to its provide technical recommendations.
practicality and ease of implementation, especially in a
rapidly digitizing market like India. Online survey tools
such as Google Forms and Survey Monkey are used to ANALYSIS & INTERPRETATION
distribute the survey, ensuring wide reach and ease of Survey & Interview Findings
participation. While convenience sampling has limitations Consumer Awareness of AI in Health Insurance
in terms of generalizability, it is suitable for exploratory
research focused on understanding consumer perspectives. The survey findings reveal that 52.9% of respondents are
aware that their insurer uses AI for claims processing, while
Demographic Breakdown The sample includes a diverse 47.1% are unaware. This indicates a near-even split in
mix of respondents in terms of age, gender, income, and type consumer awareness, with a slight majority being informed
of insurance held. The demographic breakdown is as about AI's role in claims processing. Urban respondents
follows:
(66.7% of the sample) showed higher awareness compared
● Age Group to rural respondents (33.3%), suggesting that urban areas,
o 20-30 years: 77.4% with better access to digital platforms and tech-savvy
populations, are more exposed to AI-driven systems.
o 30-40 years: 15.1%
o 40-50 years: 5.7%
o 15-20 years: 1.9%
● Gender
o Male: 62.3%
o Female: 35.8%
o Other: 1.9%
● Location
o Urban: 71.7%
o Rural: 28.3% Fig 5.1.1
● Type of Insurance Held
o Health Insurance: 64.2% ● Urban Respondents 66.7% aware of AI in claims
o Both (Health + Auto): 26.4% processing.
o Auto Insurance: 5.7% ● Rural Respondents 33.3% aware of AI in claims
processing.
o Other: 3.8%
This disparity highlights the need for insurers to improve
This demographic breakdown ensures that the study communication and education about AI adoption, especially
captures a wide range of perspectives, reflecting the in rural areas where digital literacy may be lower. One
diversity of India's insurance market.
respondent from a rural area noted, "I didn’t even know that
Limitations of the Study AI was involved in my claim process. I thought it was all
done by humans." This lack of awareness can lead to

61