aceModerate evidence
More
90 insurance shoppers trusted human-AI advice most; AI-only advice raised privacy worries
Study title: AI-driven advisory in insurance services : An experimental study on trust, transparency, and policy adoption
- Evidence type
- Randomized controlled trial
- Sample
- n = 90
- Sources
- 1 sourcePeer-reviewed
- Published
- 2 min read
- Truvace record
- TRV-2026-2541cite or check it
Study title: AI-driven advisory in insurance services : An experimental study on trust, transparency, and policy adoption
A randomized experiment with 90 insurance consumers compared three ways of getting advice on a standardized insurance policy selection: AI only, human only, and a hybrid of human and AI. The hybrid model produced the most trust, perceived transparency and intent to buy a policy.
Read the study abstract
Purpose: The rise of artificial intelligence (AI) in insurance advisory services has transformed how insurers advise their customers; however, empirical comparisons of alternative configurations of advisory services are limited.This study investigates the effects of AI-only, human-only, and hybrid human-AI advisory models on consumer trust, perceived transparency, perceived risk, and policy purchase intent.Design/methodology/approach: An experiment was designed to compare three advisory groups: AI only, human on…
58Gain side
How the two sides scoreClose call
56Problem side
The gain
In a randomized experiment with 90 insurance consumers, hybrid human-AI advice produced the most trust, perceived transparency and policy purchase intent of the three advisory models tested.
The problem
In the same 90-person experiment, AI-only insurance advice raised cyber security and privacy concerns among consumers, which worked against adoption.
A fixed formula sets both scores from the evidence and its sources. Reader votes don’t move them.
What the scores mean
G is the gain side, out of 100. It goes up when the gain is specific, when it measurably helped people and when the sources behind it are strong.
P is the problem side, out of 100. It goes up when the problem is specific, when it measurably hurt people and when the sources behind it are strong.
In brief
AI-only advice was seen as more efficient, but it also raised cyber security and privacy concerns, which worked against adoption. The findings matter for insurers deciding how much of the advisory role to hand to automated systems in a trust-dependent financial service.
The analysis also found that the advisory mode mediated the link between perceived transparency and perceived risk and consumer outcomes. AI literacy and prior claim experience shaped which advisory mode people preferred. The sample is small and the scenario was standardized rather than a real purchase.
Main points
- The experiment randomly assigned 90 insurance consumers to AI-only, human-only or hybrid human-AI advice.
- Participants went through a standardized insurance policy selection scenario.
- Hybrid advice scored highest on trust, perceived transparency and purchase intent.
- AI-only advice raised perceived efficiency but also cyber security and privacy concerns.
- Advisory mode mediated the relationship between perceived transparency, perceived risk and consumer outcomes.
- AI literacy and prior claim experience moderated which advisory mode people preferred.
The full story
The researchers used a between-groups randomized design, so each of the 90 insurance consumers saw only one advisory model. They measured trust, perceived transparency, perceived risk and policy purchase intent with validated scales.
Analysis combined analysis of variance, mediation and moderation tests using the PROCESS macro, with robustness checks using structural equation modelling. The authors say this gives causal evidence on the effect of AI in insurance advisory.
The results suggest AI-only advice gains perceived efficiency but loses ground on privacy and security worries, while hybrid advice holds up better. The authors point to the strategic value of hybrid human-AI systems in financial services where trust is essential.
What this doesn’t fix
The experiment involved only 90 consumers in a standardized policy selection scenario, which limits how far the results extend to real insurance purchases.
Sources
1 on file- Journal of Information and Optimization Sciencesdoi:10.47974/jios-2449
Talk it over
This record found a gain and a problem in the same place. Which matters more to you, and who might answer differently?
90 insurance shoppers trusted human-AI advice most; AI-only advice raised privacy worrieshttps://truvace.com/item/ai-driven-advisory-in-insurance-services-an-experimental-study-on-trust-transpar
This record found a gain and a problem in the same place. Which matters more to you, and who might answer differently?
The gain In a randomized experiment with 90 insurance consumers, hybrid human-AI advice produced the most trust, perceived transparency and policy purchase intent of the three advisory models tested.
The problem In the same 90-person experiment, AI-only insurance advice raised cyber security and privacy concerns among consumers, which worked against adoption.
Answer in the threadHow do we know? What can this kind of evidence show, and what can’t it?
The evidence This rests on a randomized controlled trial with a sample of 90, from Journal of Information and Optimization Sciences.
Answer in the threadWhat does it leave open? What else would you want to know before acting on it?
The record’s own caveat The experiment involved only 90 consumers in a standardized policy selection scenario, which limits how far the results extend to real insurance purchases.
Answer in the threadHave you seen AI used this way where you study, work or live? Did it go the same way?
Answer in the thread
The questions are the same on every record. The details under them are taken from this record as written.
What readers say