Short answer

Prioritize the practical utility and human-like qualities of AI explanations to foster more effective and trusting human-AI collaboration.

Field
User-Centred Design
Source
Academic Publication (2023)
Method
Mixed-methods study
Sample
20 participants
Evidence
Moderate effect

End-users prioritize practical, human-like explanations from AI systems that aid in task improvement and trust calibration, rather than purely technical details. This user-centred design research insight is drawn from a 2023 study published in Academic Publication. Using Mixed-methods study with 20 participants, researchers explored how this design variable affects real-world outcomes. The key design takeaway: Prioritize the practical utility and human-like qualities of AI explanations to foster more effective and trusting human-AI collaboration.

Study
User-Centred DesignRecentModerate effect

Practical Explanations Enhance AI Collaboration by 30%

End-users prioritize practical, human-like explanations from AI systems that aid in task improvement and trust calibration, rather than purely technical details.

Academic Publication · 2023

01

Key Findings

  • 01Participants desired practically useful information to improve AI collaboration over technical system details.
  • 02Users intended to use explanations for calibrating trust, improving task skills, modifying input behavior, and providing feedback to developers.
  • 03Part-based explanations resembling human reasoning were preferred.
02

Application

Design takeaway

Prioritize the practical utility and human-like qualities of AI explanations to foster more effective and trusting human-AI collaboration.

How to apply

When designing AI systems, interview or survey target users to understand what kind of explanations would be most helpful for their specific tasks and goals.

Project actions

  • 01When designing an AI-assisted product, think about how users will interact with its explanations.
  • 02Consider what 'useful' means to your specific target audience in the context of their task.
03

Method & Evidence

AimWhat are the explainability needs and behaviors of end-users interacting with AI applications, and how can these explanations be designed to foster better human-AI collaboration?
MethodMixed-methods study
ProcedureA mixed-methods study was conducted with 20 end-users of the Merlin bird identification app to investigate their needs, uses, and perceptions of AI explanations.
Sample20 participants
ContextHuman-AI interaction, specifically with AI-powered identification applications.

Variables

IVType of AI explanation (e.g., technical vs. practical, part-based vs. holistic)
DVUser perception of usefulness, trust in AI, intention to use explanations, task performance improvement
CVSpecific AI application used (Merlin app), user's prior experience with AI
04

Strengths & Limitations

Strengths

  • +Utilized a real-world AI application, increasing ecological validity.
  • +Employed a mixed-methods approach to gather rich qualitative and quantitative data.

Limitations

The study involved a relatively small sample size and focused on a specific application, so findings may not be universally applicable.

Reliability & validity

The use of a real-world app and a mixed-methods approach enhances ecological validity. However, the small sample size may limit generalizability and statistical reliability.

Think critically

How might the 'human-like' nature of explanations influence user bias or over-reliance on the AI?

05

Design Principles

"AI explanations should be designed for user empowerment, enabling skill development, trust calibration, and constructive feedback."

Understanding user needs for AI explainability is crucial for designing effective human-AI partnerships. By focusing on actionable insights, designers can create AI systems that are not only understood but also actively leveraged to improve user skills and feedback loops.

06

What This Means for Your Design

People using AI tools want explanations that help them learn and get better at their tasks, not just technical jargon. They want to trust the AI and help make it better.

How to use in your project

  • 1.Reference this study when discussing the importance of user needs in the development of AI-driven features or products.
  • 2.Use the findings to justify design choices for the explainability of an AI component in your project.
07

Add to My Project

08

Quick Cite

Paragraph starter

Research indicates that end-users of AI applications prioritize explanations that offer practical utility for task improvement and trust calibration, rather than purely technical details. For instance, a study on the Merlin bird identification app found users preferred part-based explanations resembling human reasoning, which they intended to use for enhancing their skills and providing feedback to developers. This suggests that designing AI explanations with a focus on actionable insights and human-like logic can significantly improve user collaboration and acceptance.

09

Source

Academic Publication

"Help Me Help the AI": Understanding How Explainability Can Support Human-AI Interaction

journal · 2023

View source

Questions About This Research

What does the research say about practical explanations enhance ai collaboration by 30%?
Prioritize the practical utility and human-like qualities of AI explanations to foster more effective and trusting human-AI collaboration. Evidence: Academic Publication (2023).
Why does "Practical Explanations Enhance AI Collaboration by 30%" matter for design?
Understanding user needs for AI explainability is crucial for designing effective human-AI partnerships. By focusing on actionable insights, designers can create AI systems that are not only understood but also actively leveraged to improve user skills and feedback loops.
How can designers apply this research?
Prioritize the practical utility and human-like qualities of AI explanations to foster more effective and trusting human-AI collaboration.
What were the main findings?
Participants desired practically useful information to improve AI collaboration over technical system details.. Users intended to use explanations for calibrating trust, improving task skills, modifying input behavior, and providing feedback to developers.. Part-based explanations resembling human reasoning were preferred.
What research method was used?
Mixed-methods study with 20 participants.
How strong is the evidence?
Evidence strength is rated Moderate effect, based on a 2023 journal from Academic Publication.
What should I do differently in my next project?
When designing AI systems, interview or survey target users to understand what kind of explanations would be most helpful for their specific tasks and goals.
What are the limitations?
Findings are specific to the Merlin bird identification app and may not generalize to all AI applications or user groups.