Short answer
Designers should proactively seek to understand the specific 'why' and 'how' questions users have about AI systems, and advocate for XAI features that directly address these queries, rather than relying solely on algorithmically generated explanations.
- Field
- Innovation & Design
- Source
- Academic Publication (2020)
- Method
- Qualitative research through expert interviews and the development of a question bank.
- Sample
- 20 participants
- Evidence
- Moderate effect
Design practitioners identify a significant disconnect between the technical advancements in explainable AI (XAI) algorithms and the practical needs of users for understanding AI systems. This innovation & design research insight is drawn from a 2020 study published in Academic Publication. Using Qualitative research through expert interviews and the development of a question bank. with 20 participants, researchers explored how this design variable affects real-world outcomes. The key design takeaway: Designers should proactively seek to understand the specific 'why' and 'how' questions users have about AI systems, and advocate for XAI features that directly address these queries, rather than relying solely on algorithmically generated explanations.
Bridging the Gap: Translating Algorithmic Explainability into User-Centric AI Design
Design practitioners identify a significant disconnect between the technical advancements in explainable AI (XAI) algorithms and the practical needs of users for understanding AI systems.
Academic Publication · 2020
Key Findings
- 01There is a notable divergence between the focus of current XAI algorithmic research and the practical requirements for user understanding in AI products.
- 02Design practitioners require tools and frameworks that translate complex algorithmic explanations into actionable design strategies for user interfaces.
- 03User needs for AI explainability are best articulated through the types of questions users would naturally ask about an AI's behavior and decision-making.
Application
Design takeaway
Designers should proactively seek to understand the specific 'why' and 'how' questions users have about AI systems, and advocate for XAI features that directly address these queries, rather than relying solely on algorithmically generated explanations.
How to apply
When designing AI-powered features, create a 'user question bank' based on anticipated user inquiries about the AI's functionality, decision-making, and potential biases. Use this bank to inform the design of explainability interfaces.
Project actions
- 01When designing an AI system, consider what questions a user might have about its outputs or decisions.
- 02Explore how to visually represent or verbally communicate the reasoning behind an AI's actions in a clear and concise manner.
Method & Evidence
Variables
Strengths & Limitations
Strengths
- +Focuses on the practical application of XAI in design.
- +Employs qualitative methods to capture nuanced practitioner perspectives.
Limitations
The number of practitioners interviewed is relatively small, and the types of AI products they work on might influence their responses.
Reliability & validity
The validity of the findings relies on the rich qualitative data from interviews and the structured approach of the question bank. Reliability could be enhanced by using multiple researchers for coding interview data.
Think critically
How can designers effectively anticipate and address the diverse range of user questions about AI, especially when the AI's internal workings are highly complex or proprietary?
Design Principles
"Translate algorithmic complexity into user-understandable explanations by anticipating and addressing user-generated questions."
This gap highlights a critical area for innovation in AI product development. Designers and engineers must move beyond purely algorithmic solutions to create AI systems that are not only functional but also transparent and trustworthy for end-users.
What This Means for Your Design
Even though AI can explain itself better now, designers aren't sure how to show that explanation to people in a way that makes sense. This research found that designers need to think about the questions users would actually ask the AI to make explanations useful.
How to use in your project
- 1.Use the concept of a 'user question bank' to justify your design choices for AI explainability features.
- 2.Reference the identified gap between algorithmic XAI and user needs to highlight the importance of your user-centered design approach.
Add to My Project
Quick Cite
Paragraph starter
This design project addresses the critical challenge of translating technical explainable AI (XAI) capabilities into user-centered experiences. Research indicates a significant gap between algorithmic advancements in XAI and the practical needs of users for understanding AI systems. By focusing on the types of questions users would naturally ask about an AI, this project aims to bridge this divide and develop more transparent and trustworthy AI products.
Source
Academic Publication
Questioning the AI: Informing Design Practices for Explainable AI User Experiences
journal · 2020
View sourceQuestions About This Research
- What does the research say about bridging the gap: translating algorithmic explainability into user-centric ai design?
- Designers should proactively seek to understand the specific 'why' and 'how' questions users have about AI systems, and advocate for XAI features that directly address these queries, rather than relying solely on algorithmically generated explanations. Evidence: Academic Publication (2020).
- Why does "Bridging the Gap: Translating Algorithmic Explainability into User-Centric AI Design" matter for design?
- This gap highlights a critical area for innovation in AI product development. Designers and engineers must move beyond purely algorithmic solutions to create AI systems that are not only functional but also transparent and trustworthy for end-users.
- How can designers apply this research?
- Designers should proactively seek to understand the specific 'why' and 'how' questions users have about AI systems, and advocate for XAI features that directly address these queries, rather than relying solely on algorithmically generated explanations.
- What were the main findings?
- There is a notable divergence between the focus of current XAI algorithmic research and the practical requirements for user understanding in AI products.. Design practitioners require tools and frameworks that translate complex algorithmic explanations into actionable design strategies for user interfaces.. User needs for AI explainability are best articulated through the types of questions users would naturally ask about an AI's behavior and decision-making.
- What research method was used?
- Qualitative research through expert interviews and the development of a question bank. with 20 participants.
- How strong is the evidence?
- Evidence strength is rated Moderate effect, based on a 2020 journal from Academic Publication.
- What should I do differently in my next project?
- When designing AI-powered features, create a 'user question bank' based on anticipated user inquiries about the AI's functionality, decision-making, and potential biases. Use this bank to inform the design of explainability interfaces.
- What are the limitations?
- The study's findings are based on the perspectives of a specific group of design practitioners and may not represent all user needs or all AI development contexts.