GenAI Fundamentals for UX Designers + Researchers
Learn the key concepts and components to harness GenAI for UX design — and lead AI product innovation efforts.
4.43 (568 reviews)
2,824
students
3 hours
content
Oct 2024
last update
$49.99
regular price
What you will learn
Learn core principles of responsible GenAI design that ensure our products serve all users equally + equitably.
Success factors unique to Machine Learning (ML) products that UX + Product Designers and their teams should adopt
Guidance on research to determine what kinds of problems are best solved by AI, and where human control should remain central.
Determining when AI features are appropriate for users — and when they aren't
Identifying when Automation (AI does the task for users) or Augmentation (help them do it better) is more appropriate.
Designing the reward function and appropriately considering the balance between false positives and false negatives.
Essential factors to consider when evaluating the reward function.
How to weigh necessary, unavoidable tradeoffs between precision and recall, which is key to shaping an AI user experience.
Designing for fairness and inclusion, from objectives to datasets to guidance on bias testing.
Designing for generative variability: how do we present multiple, varied outputs to users — and how do we guide them in selecting the best one?
Designing for multiple outputs: how do we help users filter and highlight differences between outputs?
Designing for imperfections: how do we empower users to manage + mitigate imperfections and designing with contextual sensitivity?
Designing for confidence: how do we design confidence scoring to properly evaluate output quality and increase user trust?
Rules and examples for applying confidence scores.
Designing for co-creation: how to we design co-creation processes where both the user and the AI can make adjustments?
Designing for generic controls: using "temperature" to control the number of outputs and the degree of variability in those outputs.
Designing for domain-specific controls, such as encoder-decoder models, semantic sliders and prompt engineering.
Designing for prompt engineering: enabling and guiding users to effectively use multiple types of conversational prompts.
Designing for exploration: incorporating flexibility, feedback, transparency and error handling/expectation management.
Designing for choice, feedback, transparency + safety: centering users as active, empowered participants in the creation process.
Designing for mental models: orienting users to generative variability, teaching effective use and teaching the AI about the user.
Designing for explanation, understanding + trust: providing clear rationales for outputs, using friction to curb over-reliance and showing imperfections.
Designing against harm: understanding the critical ways irresponsible AI design can harm people.
Designing against hazardous outputs: discrimination, exclusion, toxicity, misinformation, deep fakes, IP theft and more.
Mitigating harm with a Value-Sensitive Design (VSD) process, integrated with an Agile or Lean development process.
6127575
udemy ID
8/14/2024
course created date
10/29/2024
course indexed date
Bot
course submited by