40 min
#143 - Harnessing AI For Better Insights with George Whitfield of MIT and FindOurView Awkward Silences
-
- Technology
Carol and Erin welcome George Whitfield, an expert in applying AI to the analysis of qualitative data. George discusses the intricate challenges of leveraging language models to interpret expansive open-ended data (like interview transcripts), emphasizing the importance of context and not just keyword or topic identification.
They'll dig into the crucial role of human oversight in AI, what preliminary analysis might look like using AI, how to check and refine the work of an AI assistant without derailing your project delivery date, and recommendations for etiquette regarding the reporting of AI-informed results.
The episode closes with an exploration of the limits of AI and where user experience researchers can play a larger role in its development. George believes AI can (and should) inspire new directions of research, but not dictate them.
Episode Highlights
03:48 - Innovating consumer insights using AI12:21 - Importance of human involvement in AI tools20:04 - Enhance discussion sections with AI tools26:50 - AI-inspired insights provide inspiration, not guidance34:12 - Interpretation beyond analyzing transcripts36:46 - Applying engineering rigor to the process of building a businessAbout Our GuestGeorge Whitfield is an Entrepreneur-in-Residence at the Martin Trust Center for MIT Entrepreneurship, Lecturer at the MIT Sloan School of Management, and CEO at FindOurView. As CEO of his most recent company FindOurView, he launched a Gen AI product to help user researchers synthesize insights faster from high volumes of customer interviews. George holds 4 patents and has 3 degrees from MIT including a Bachelors in Electrical Engineering and Computer Science and a Masters and Ph.D. in Materials Science and Engineering.
Resources on Qual Data and AI
Our AI in UX Report shares findings from a survey of over 1,000 researchersWhat does it mean to "code" qualitative data? This breakdown explains it all.Interested in trying an AI tool for your analysis? Here are 20 worth considering.
Carol and Erin welcome George Whitfield, an expert in applying AI to the analysis of qualitative data. George discusses the intricate challenges of leveraging language models to interpret expansive open-ended data (like interview transcripts), emphasizing the importance of context and not just keyword or topic identification.
They'll dig into the crucial role of human oversight in AI, what preliminary analysis might look like using AI, how to check and refine the work of an AI assistant without derailing your project delivery date, and recommendations for etiquette regarding the reporting of AI-informed results.
The episode closes with an exploration of the limits of AI and where user experience researchers can play a larger role in its development. George believes AI can (and should) inspire new directions of research, but not dictate them.
Episode Highlights
03:48 - Innovating consumer insights using AI12:21 - Importance of human involvement in AI tools20:04 - Enhance discussion sections with AI tools26:50 - AI-inspired insights provide inspiration, not guidance34:12 - Interpretation beyond analyzing transcripts36:46 - Applying engineering rigor to the process of building a businessAbout Our GuestGeorge Whitfield is an Entrepreneur-in-Residence at the Martin Trust Center for MIT Entrepreneurship, Lecturer at the MIT Sloan School of Management, and CEO at FindOurView. As CEO of his most recent company FindOurView, he launched a Gen AI product to help user researchers synthesize insights faster from high volumes of customer interviews. George holds 4 patents and has 3 degrees from MIT including a Bachelors in Electrical Engineering and Computer Science and a Masters and Ph.D. in Materials Science and Engineering.
Resources on Qual Data and AI
Our AI in UX Report shares findings from a survey of over 1,000 researchersWhat does it mean to "code" qualitative data? This breakdown explains it all.Interested in trying an AI tool for your analysis? Here are 20 worth considering.
40 min