Short answer
Adopt a more adaptive and iterative design process that anticipates and plans for AI's inherent unpredictability and complexity, rather than assuming deterministic behavior.
- Field
- User-Centred Design
- Source
- Figshare (2021)
- Method
- Qualitative research, synthesis of existing literature, industry case studies, and teaching observations.
- Evidence
- Strong effect
The inherent uncertainty in AI capabilities and the complexity of its outputs present unique challenges for user experience design, requiring designers to develop new methods beyond traditional approaches. This user-centred design research insight is drawn from a 2021 study published in Figshare. Using Qualitative research, synthesis of existing literature, industry case studies, and teaching observations., researchers explored how this design variable affects real-world outcomes. The key design takeaway: Adopt a more adaptive and iterative design process that anticipates and plans for AI's inherent unpredictability and complexity, rather than assuming deterministic behavior.
AI's Unpredictability Demands Novel UX Design Strategies
The inherent uncertainty in AI capabilities and the complexity of its outputs present unique challenges for user experience design, requiring designers to develop new methods beyond traditional approaches.
Figshare · 2021
Key Findings
- 01The primary UX design challenges for AI stem from uncertainty in AI system capabilities.
- 02The complexity of potential AI system outputs is a significant design hurdle.
- 03Existing UX design methods may not fully address the unique failure modes of AI.
Application
Design takeaway
Adopt a more adaptive and iterative design process that anticipates and plans for AI's inherent unpredictability and complexity, rather than assuming deterministic behavior.
How to apply
When designing an AI-powered feature, explicitly map out potential AI errors and unpredictable outputs. Develop user interfaces that clearly communicate AI confidence levels or potential inaccuracies, and provide mechanisms for user feedback and correction.
Project actions
- 01When designing with AI, think about what could go wrong and how users will react.
- 02Consider how to show users when the AI might be wrong or uncertain.
Method & Evidence
Variables
Strengths & Limitations
Strengths
- +Addresses a timely and critical issue in modern design practice.
- +Synthesizes diverse sources of knowledge to build a comprehensive understanding.
Limitations
It can be difficult to fully simulate the unpredictable nature of real-world AI in a controlled project setting.
Reliability & validity
Reliability could be enhanced by using standardized metrics for user experience and having multiple researchers assess qualitative data. Validity is supported by drawing on multiple research methodologies and industry insights.
Think critically
How can designers proactively build trust in AI systems when their behavior is inherently uncertain and can lead to unpredictable errors?
Design Principles
"Design for uncertainty: Embrace and manage the inherent unpredictability of AI systems by building in flexibility, transparency, and robust error handling."
As AI becomes more integrated into everyday products and services, understanding and addressing its unique design challenges is crucial for creating effective, reliable, and trustworthy user experiences. Designers must move beyond mitigating standard technical failures to account for the nuanced and often unpredictable nature of AI.
What This Means for Your Design
AI can be tricky to design for because it doesn't always work perfectly or predictably. Designers need new ways to handle these unexpected problems to make AI products user-friendly.
How to use in your project
- 1.Use the identified challenges of AI uncertainty and output complexity to justify your design choices and the methods you employ in your design project.
Add to My Project
Quick Cite
Paragraph starter
The design of AI systems presents unique challenges due to inherent uncertainty in their capabilities and the complexity of their outputs. This necessitates a departure from traditional UX design approaches, requiring methods that can effectively manage user expectations and mitigate unintended consequences arising from AI's probabilistic nature and potential for unexpected behaviors.
Source
Figshare
Profiling Artificial Intelligence as a Material for User Experience Design
journal · 2021
View sourceQuestions About This Research
- What does the research say about ai's unpredictability demands novel ux design strategies?
- Adopt a more adaptive and iterative design process that anticipates and plans for AI's inherent unpredictability and complexity, rather than assuming deterministic behavior. Evidence: Figshare (2021).
- Why does "AI's Unpredictability Demands Novel UX Design Strategies" matter for design?
- As AI becomes more integrated into everyday products and services, understanding and addressing its unique design challenges is crucial for creating effective, reliable, and trustworthy user experiences. Designers must move beyond mitigating standard technical failures to account for the nuanced and often unpredictable nature of AI.
- How can designers apply this research?
- Adopt a more adaptive and iterative design process that anticipates and plans for AI's inherent unpredictability and complexity, rather than assuming deterministic behavior.
- What were the main findings?
- The primary UX design challenges for AI stem from uncertainty in AI system capabilities.. The complexity of potential AI system outputs is a significant design hurdle.. Existing UX design methods may not fully address the unique failure modes of AI.
- What research method was used?
- Qualitative research, synthesis of existing literature, industry case studies, and teaching observations..
- How strong is the evidence?
- Evidence strength is rated Strong effect, based on a 2021 journal from Figshare.
- What should I do differently in my next project?
- When designing an AI-powered feature, explicitly map out potential AI errors and unpredictable outputs. Develop user interfaces that clearly communicate AI confidence levels or potential inaccuracies, and provide mechanisms for user feedback and correction.
- What are the limitations?
- The framework may need further validation across a broader range of AI applications and industries. The specific methods for addressing these challenges are still emerging.