AI-driven systems increasingly provide recommendations across domains such as finance, health, and productivity. As these systems become embedded in everyday decision-making, the way information is framed may influence not only trust, but also perceived autonomy and compliance. While accuracy and performance are widely studied, the psychological impact of tone framing remains comparatively underexplored.
Context
While AI recommendation systems are designed to assist users, the framing of suggestions may subtly shift perceived authority and autonomy. An authoritative tone could increase compliance but reduce perceived agency, whereas probabilistic language might preserve autonomy but weaken trust. Understanding this balance is critical for designing responsible AI interactions.
Problem Space
How do different AI tone framings (authoritative, neutral, probabilistic) influence user trust, perceived agency and behavioral decision shift?
Research Question
12 PARTICIPANTS (N = 12)
RANDOM ASSIGNMENT
BETWEEN-SUBJECT DESIGN
CONDITION A
Authoritative
(n = 4)
Neutral
CONDITION B
(n = 4)
MEASURES
POST-EXPOSURE
Probabilistic
CONDITION C
(n = 4)
Trust · Agency · Influence · Reliance · Decision Shift
Procedure
01
02
Participants were randomly assigned to one of three tone conditions (authoritative, neutral, probabilistic).
Each participant was asked to choose between two options: Project Alpha and Project Beta.
After making an initial selection, an AI-generated recommendation appeared suggesting the opposite option to the one previously chosen.
The recommendation was framed using the assigned tone condition.
03
04
Participants were then given the opportunity to either:
・Confirm their initial decision
・Change their choice in response to the AI suggestion
05
06
After the decision phase, participants completed a short questionnaire measuring:
・Trust ・Perceived agency ・Influence ・Reliance ・Decision shift
Analysis
Data were analyzed by comparing the three tone conditions:
・Authoritative, Neutral, Probabilistic
・Mean scores calculated for trust, agency, influence
reliance
・Decision shift rate measured as % of participants
changing their initial choice.
Analysis Overview
Authoritative
Neutral
Probabilistic
75%
0%
25%
Decision Shift Rate
The percentage of participants who changed their initial decision was calculated for each condition:
・Authoritative: 75% (3 out of 4
partecipants)
・Neutral: 0% (0 out of 4 partecipants)
・Probabilistic: 25% (1 out of 4
partecipants)
The authoritative tone produced the highest decision shift rate, suggesting stronger behavioral influence compared to the other framing styles.
Mean Score Comparison
Perceived Agency
· Probabilistic tone scored highest
· Neutral tone maintained stable trust
· Authoritative tone had highest influence
· Higher reliance under Authoritative
Trust
Influence
Reliance
Key Insight
The authoritative tone maximizes behavioral compliance but reduces perceived trust. This reveals a fundamental design tension between persuasion effectiveness and user credibility perception
Design Implications
Limitations
·
The study was conducted with a small sample size (n=12), limiting statistical generalizability
·
Participants were exposed to a simulated decision-making scenario rather than a real-world AI system
·
The interaction measured short-term behavioral response without longitudinal trust tracking
Future Directions
·
Conduct a larger-scale study to validate tone-related behavioral effects across diverse user group
·
Implement real-world A/B testing within AI-driven interfaces to measure behavioral and trust outcomes at scale
·
Explore adaptive tone systems that dynamically adjust framing based on context and user history
·
Investigate long-term trust development through longitudinal interaction tracking
Research Reflection
This project reinforced how subtle interface framing can meaningfully influence user autonomy perception. As a researcher, I learned the importance of balancing behavioral effectiveness with long-term trust preservation.
Translating behavioral findings into product-level decisions.
SALVATORE
FANILE
TECHNOLOGY
TECHNOLOGY
USER RESEARCH FOCUSED ON HUMAN - INTERACTION
I explore how digital systems influence human perception, behavior, and decision making.
My work combines psychological research and emerging technologies to design more human-centered AI experiences.
View Work
Authoritative Tone
Authoritative Tone
increases compliance
but reduces perceived
agency
Use in high-risk or safety-critical systems
Avoid in autonomy-sensitive tools (education, therapy, creative platforms)
·
·
·
·
Neutral Tone
Neutral framing
maintains trust stability
·
·
Ideal for transactional systems
Suitable for long-term user relationship
Probabilistic Tone
Probabilistic framing
supports perceived
autonomy
·
·
Use in decision-support tools
Enhances collaborative AI perception
·
Contact
© 2026 Salvatore Fanile
Available for
UX Research roles.
Bridging psychology
and technology through
user research.
Hello, I'm Salvatore
My work sits at the intersection of cognitive psychology and intelligent systems.
I study how digital environments influence trust, agency, and decision-making.
With a background in psychology and a focus on emerging technologies, I explore how AI communication and system design shape human behavior.
I aim to contribute to more ethical, human-centered technological systems through research and experimentation.
About
Work
The Behavioral Impact of AI Tone
This experimental study investigates how different AI tone framings (authoritative, neutral, probabilistic) influence user trust, perceived agency, and behavioral decision shift.
Results show significant variation in compliance and autonomy perception across tone conditions.
ROLE
UX Researcher
STUDY TYPE
Experimental Study
Between-subject design
12 partecipants
Trust · Agency · Influence · Reliance · Decision Shift
3 weeks
DESIGN
PARTECIPANTS
MEASURES
DURATION