Short answer
Prioritize designing for human-to-human trust and communication around AI systems, as this often has a greater impact on user trust than direct AI system features.
- Field
- Human Factors
- Source
- Academic Publication (2024)
- Method
- Qualitative research through semi-structured interviews
- Sample
- 14 participants (7 AI practitioners, 7 decision subjects)
- Evidence
- Strong effect
Trust in AI-assisted decision-making is more significantly shaped by the perceptions and actions of other human actors involved than by the inherent features of the AI system itself. This human factors research insight is drawn from a 2024 study published in Academic Publication. Using Qualitative research through semi-structured interviews with 14 participants (7 AI practitioners, 7 decision subjects), researchers explored how this design variable affects real-world outcomes. The key design takeaway: Prioritize designing for human-to-human trust and communication around AI systems, as this often has a greater impact on user trust than direct AI system features.
Human influence on AI trust outweighs system design
Trust in AI-assisted decision-making is more significantly shaped by the perceptions and actions of other human actors involved than by the inherent features of the AI system itself.
Academic Publication · 2024
Key Findings
- 01Participants could identify prerequisites for trust and differentiate it from related concepts like trustworthiness, reliance, and compliance.
- 02Trust in AI systems is more heavily influenced by other human actors than by the system's own features.
- 03The factors contributing to human-AI trust are dependent on the specific stakeholder group.
Application
Design takeaway
Prioritize designing for human-to-human trust and communication around AI systems, as this often has a greater impact on user trust than direct AI system features.
How to apply
When designing AI-assisted decision systems, map out all human actors involved and understand their relationships and communication flows. Design interventions that build trust between these human actors, in addition to designing the AI interface.
Project actions
- 01When researching user trust in a design, consider interviewing not just the end-user but also anyone who supports, maintains, or is involved in the decision-making process with the user.
- 02Think about how communication between different stakeholders can be designed to build or erode trust in a system.
Method & Evidence
Variables
Strengths & Limitations
Strengths
- +Explores the under-researched perspectives of both AI practitioners and decision subjects.
- +Differentiates trust from related concepts, providing a more nuanced understanding.
Limitations
The findings are based on qualitative interviews and may be subjective. The specific AI systems and decision domains were not detailed, making it hard to generalize specific design recommendations.
Reliability & validity
The qualitative nature of interviews provides rich data but may have lower inter-rater reliability. Validity is strengthened by exploring diverse perspectives from different stakeholder groups.
Think critically
If human actors are more influential than system features in building trust, how can designers ethically leverage or mitigate this influence, especially when human actors might have biases or incomplete understanding?
Design Principles
"Human-centric AI design must account for the social network and interpersonal dynamics influencing user trust."
This insight challenges the common assumption that improving AI algorithms or user interfaces is the primary driver of trust. It highlights the critical need for designers and developers to consider the broader social and interpersonal dynamics surrounding AI deployment.
What This Means for Your Design
When people decide whether to trust an AI, they care more about what other people say and do about the AI than about the AI itself. Who these other people are also matters.
How to use in your project
- 1.Use this research to justify exploring the social context and stakeholder relationships in your design project, rather than solely focusing on the technical aspects of your proposed solution.
- 2.Cite this study when discussing how user trust is influenced by factors beyond the product's direct features.
Add to My Project
Quick Cite
Paragraph starter
This research indicates that trust in AI-assisted decision-making is significantly influenced by human actors and social dynamics, rather than solely by the AI system's features. Therefore, a comprehensive design approach must consider the interpersonal relationships and communication pathways among all stakeholders involved in the AI's deployment and use.
Source
Academic Publication
Trust in AI-assisted Decision Making: Perspectives from Those Behind the System and Those for Whom the Decision is Made
journal · 2024
View sourceQuestions About This Research
- What does the research say about human influence on ai trust outweighs system design?
- Prioritize designing for human-to-human trust and communication around AI systems, as this often has a greater impact on user trust than direct AI system features. Evidence: Academic Publication (2024).
- Why does "Human influence on AI trust outweighs system design" matter for design?
- This insight challenges the common assumption that improving AI algorithms or user interfaces is the primary driver of trust. It highlights the critical need for designers and developers to consider the broader social and interpersonal dynamics surrounding AI deployment.
- How can designers apply this research?
- Prioritize designing for human-to-human trust and communication around AI systems, as this often has a greater impact on user trust than direct AI system features.
- What were the main findings?
- Participants could identify prerequisites for trust and differentiate it from related concepts like trustworthiness, reliance, and compliance.. Trust in AI systems is more heavily influenced by other human actors than by the system's own features.. The factors contributing to human-AI trust are dependent on the specific stakeholder group.
- What research method was used?
- Qualitative research through semi-structured interviews with 14 participants (7 AI practitioners, 7 decision subjects).
- How strong is the evidence?
- Evidence strength is rated Strong effect, based on a 2024 journal from Academic Publication.
- What should I do differently in my next project?
- When designing AI-assisted decision systems, map out all human actors involved and understand their relationships and communication flows. Design interventions that build trust between these human actors, in addition to designing the AI interface.
- What are the limitations?
- The study involved a small sample size and may not be generalizable to all AI decision-making contexts or diverse user populations. The specific domains of decision-making were varied, which could introduce confounding factors.