Key resources to define UX for AI products.
Foundations, fundamentals, basics and principles that you need to consider when designing for AI.
One of the essential aspects of designing for AI is to have some common ground on how to create the best user experience for your product in AI. By setting expectations, starting small, being pragmatic, using it in the right context, and being transparent, your user will embrace your AI product or integration.
Here are ten five resources for UX and design focused on creating products with AI, along with personal takeaways of their usefulness.
UX for AI
It is a comprehensive resource focused on designing user-centered AI products. It outlines core design principles essential for creating effective and humane AI technologies. Key topics include setting user expectations, explaining AI results, communicating algorithm confidence, and maintaining user control.
Personal takeaway/ Build trust over time. Everything new in your product is scary and changes your user’s perception. Ship things gradually, so that the system can be more intelligent and your user can get used to the intelligence.
IBM Design for AI
IBM Design for AI Fundamentals provides essential guidelines for designing AI products. It covers key aspects such as AI design foundations, characteristics of AI, design factors, and relationship development between AI and users.
Personal takeaway/ Nothing really. Why not? Because all of their fundamentals are very accurate (the transaction age vs transaction age), but very dysfunctional and philosophical. For design purposes, this means that it’s hard to apply and imagine how to use it as design fundamentals when wanting to integrate AI into products.
MutualMarketing human dynamics overview. Builds on Debra M. Amidon’s ‘Evolution of Thought’ (IBM Design)
Predictably Smart by Google Design
The article “Predictably Smart” on Google Design discusses the integration of machine learning (ML) in user interfaces. It emphasizes that while ML-driven recommendations and personalization can enhance the user experience by saving time and effort, they can also disrupt the user’s habituation process if not implemented thoughtfully. The article provides principles for balancing ML features with predictable UI design, such as counting evaluation steps as navigation steps, maintaining predictable UI for high-stakes tasks, dedicating specific areas for ML features, and planning for ML failures.
Personal takeaway/ Make failure your baseline. Set the standard that your algorithm is making bad predictions and see what the user’s process will be without and with the assistance of the AI. If it creates more effort to complete with the assistance… you know the answer.
People + AI Guidebook
The “Patterns” section of the Google PAIR Guidebook offers practical design patterns for integrating AI into user experiences. It provides specific guidelines and best practices for various aspects of AI design, such as handling errors, offering explanations, and managing user expectations. These patterns help designers create intuitive and effective AI-driven products.
Personal takeaway/ Add context from human sources. Explain why and where your data is found. Second, show, don’t tell. Users want to understand where your data is coming from, that it’s defined by research done by equals, humans, not computers.
Adding social proof to recommendations will strengthen the value of the data.
IDEO and Healthworx Venture Studio collaborated to explore how to design new ventures by integrating human-centered design and generative AI. They found that AI can significantly increase efficiency in tasks like research, ideation, and prototyping, but human input remains crucial for nuanced insights, creativity, and innovative solutions. They developed a “Scale of AI Utility” framework categorizing AI applications as novel, transformative, inspirational, or unreliable for different design tasks. While AI struggles with contextual awareness and subject matter expertise, leveraging both human and AI capabilities through an iterative process can optimize the innovation lifecycle. The future involves a symbiotic human-AI partnership where AI augments and accelerates human ingenuity.
Personal takeaway/ Distinct between unique values for humans and for AI, so you don’t mix what humans value (like gut feeling) and what an AI can bring (more efficiency, less time). So a tool that will be more efficient, isn’t always more intuitive for the audience you serving.
These resources are valuable because they cover a wide spectrum of essential topics, including human-centered design principles, ethical considerations, practical implementation strategies, and real-world case studies, all crucial for creating successful AI products.
👾 Interested in more content around Product Design for AI and awesome tools that can help you build, scale & iterate your product?
→ Follow me on X / Twitter
→ Tools & templates for product managers in my shop
👾 Product by Pim is free today. But if you enjoyed this post, you can always support me via my Buymeacoffee
