Summary
Join Savina Hawkins as she presents a focused exploration of large language models in User Experience Research (UXR). This talk zeroes in on practical, actionable strategies for leveraging these advanced AI tools to enhance UXR outcomes. Savina will discuss the dual nature of large language models, illuminating their potential to both revolutionize and challenge UXR practices. Attendees will learn to identify and mitigate risks while maximizing the benefits of these models in real-world scenarios. The session promises a clear, concise roadmap for UX researchers to effectively integrate large language models into their workflow, ensuring positive, impactful results.
Key Insights
-
•
Judd Anton predicts prompt engineering will be a foundational skill for researchers by 2025.
-
•
AI tools have enabled condensing thousands of user insights into manageable unique ideas within minutes.
-
•
Large language models learn language skills purely from statistical patterns, not human understanding.
-
•
GPT’s transformer architecture allows fast and contextually relevant text generation.
-
•
Prompt engineering is both a science of optimization and an art of design mirroring qualitative interview skills.
-
•
Prompt chaining can sequentially automate complex, multi-step analysis workflows.
-
•
AI-generated outputs are probabilistic and sometimes incorrect, requiring critical human oversight.
-
•
Multi-threaded prompt chaining workflows can achieve human-quality qualitative analysis at scale.
-
•
The rise of AI in knowledge work challenges existing power structures, workflows, and capitalism itself.
-
•
Researchers’ humanistic skills uniquely position them to co-design ethical AI systems and maintain meaning-making.
Notable Quotes
"The best researchers are multi-method and they have five tools, with prompt engineering as the fifth emerging one."
"In just 15 minutes, I extracted thousands of quotes and turned feedback into a strategic opportunity map in 30 minutes."
"GPT is like a master chef who’s traveled the world sampling cuisines before creating new dishes."
"Large language models don’t understand language like we do; they generate statistically likely text based on training data."
"Prompt engineering is the craft of using specific inputs to elicit precise AI outputs, similar to designing discussion guides."
"Chains of prompts become designed information processing workflows that can scale research tasks."
"These AI tools can do the entire qualitative analysis process at the click of a button."
"Magic eight ball thinking is the risk of blindly accepting AI answers without critical evaluation."
"Artificial intelligence can’t produce meaning the way humans do; it industrializes plausible content generation."
"Our challenge is to design how humans and technology share the world in peace while shaping new knowledge work."
Or choose a question:
More Videos
"I was trained to isolate very specific data points and only assert what was backed by evidence."
Michele MarutResearch Repositories Reconsidered
February 14, 2019
"Snow clearing policies choose to clear roads before pavements, deeply ignoring responsibilities that are likely to fall on women."
Mansi GuptaDrawing from Feminist Practice to Make Inclusive Design Operational
September 9, 2022
"Scott Cook noticed his wife’s frustration balancing their checkbook and realized millions more probably felt the same pain."
Kaaren HansonStop Talking, Start Doing
June 9, 2017
"Correlation doesn’t imply causation, but sometimes two things happening together do imply a causal mechanism."
Joshua NobleCasual Inference
October 6, 2023
"A typical enterprise customer has like five analysts who need a login, not millions of users."
Dan WillisEnterprise Storytelling Sessions
May 13, 2015
"Turning every study plan into a kind of report with new and existing insights is a mindset shift that can build a durable asset."
Jake BurghardtFinding More Inroads into Research Impact
February 20, 2026
"Our ability to entice participants to share their most personal stories raises the potential to exploit them for design research."
Rachael Dietkus, LCSWTrauma-Responsive Design: Reimagining the Future of Design Now
December 10, 2021
"Put the most important message right up front no matter what the format."
Megan BlockerGetting to the “So What?”: How Management Consulting Practices Can Transform Your Approach to Research
March 26, 2024
"Auto captions on YouTube struggle with accents and technical terms, so manual or human-reviewed captioning is essential for accuracy."
Sheri Byrne-HaberThe Importance of Accessible Design Systems
January 8, 2024
