Designing AI to Make Users Smarter
September 23
9:00a - 5:00p ET
Location TBD
Christopher Noessel
Author & Design Principal
About this workshop
AI is full of promise for users, but it introduces risk as well. The two we’ll talk about in this session are over-reliance and deskilling. Over-reliance is when users trust an AI’s output too much. Deskilling is when users lose skills they previously had, but handed off to the AI; with implications not just for users but for labor relations as well. The good news is that you as a designer can do something about each of these. Come hear Christopher Noessel introduce the problems, share examples, and walk us through the patterns we can implement to help take some of the sting out of AI.
Target Audience
<p>Designers and Product managers working with AI-infused software meant to help users perform tasks, or who expect to be in the future.</p>
Pre-requisites
<p>Familiarity with user-centered design practice. Passing familiarity with modern AI services, apps, and software.</p>
Takeaways
- How software agents are different from software assistants, along with clarifying examples.
- The AI technologies that enable universal assists.
- The Universal Assists framework, a tool for reviewing, designing, and critiquing assistants.
- The risks that assistance introduces: overreliance and deskilling.
- Tools to mitigate the risks.
- What cognitive styles contribute to the risk of overreliance.
High-Level Agenda
Day 1
- Getting crisp on assistants
- Exercise and discussion
- Break
- The see-think-do loop and the 5 Universal Assists
- Exercises and discussions
Day 2
- Overreliance and Deskilling
- Exercise and discussion
- Cognitive forcing functions and the Human Goes First pattern
- Exercise and discussion
- Break
- Cognitive styles and the cognitive categorizer study
- Exercise and discussion
- Q&A (as time allows)
Session speaker