AI is making it easier than ever to run research, but faster doesn't always mean better. In this episode, we dig into what it really means to democratize research responsibly, and why your team probably needs a charter before someone does something they can't take back.

Your team is already running research without you. So the real question is: are you going to help them do it well, or just hope for the best?

Ned Dwyer is the co-founder and CEO of Great Question, an all-in-one UX research platform built to bring research to everyone in an organization. Not just the people with "researcher" in their title. He's spent years thinking about how teams can democratize access to customer insights without turning research into a free-for-all, and his talk at UX Con is what first put him on my radar.

In this conversation, we dig into one of the more divisive topics in our industry right now: research democratization. Ned makes a pretty compelling case that it's not the all-or-nothing argument a lot of people make it out to be. It's a spectrum, and where your organization should land on that spectrum depends on who you're researching, what decisions are being made, and how much risk is on the table. We also get into AI's role in all of this, from AI-moderated interviews to synthesized insights, and where teams tend to get themselves into trouble when they hand over too much to the machine without any real governance in place.

The thing I found most useful in this conversation is Ned's concept of a democratization charter, a practical framework for defining who should be doing what kind of research, with which populations, and under what guardrails. It's something I honestly hadn't thought much about before meeting Ned, and I think it's one of the most actionable ideas we've talked about on the show. If your team is already using AI research tools (and let's be honest, they probably are), this conversation is worth your time.

Topics:

• 01:45 - Ned's origin story and why he built Great Question

• 04:10 - The pressure to move fast, and what gets lost when speed wins

• 06:11 - The 80/20 rule: how to use AI without publishing slop

• 09:45 - Democratization is a spectrum, not a binary

• 12:35 - Where guardrails matter most: vulnerable populations and one-way-door decisions

• 13:12 - The case for a democratization charter

• 19:00 - AI moderation demystified: closer to a talking survey than a human interviewer

• 23:00 - Ned's GoDaddy confession: how rogue research goes wrong

• 27:00 - Participant fatigue and insight overload: the new risks AI introduces

• 31:45 - Rogue research will happen regardless... your job is to make it safer

• 43:28 - The Will Smith spaghetti analogy and where AI tools are headed

—

Thanks for listening! We hope you dug today’s episode. If you liked what you heard, be sure to like and subscribe wherever you listen to podcasts! And if you really enjoyed today’s episode, why don’t you leave a five-star review? Or tell some friends! It will help us out a ton.

If you haven’t already, sign up for our email list. We won’t spam you. Pinky swear.

• ⁠⁠⁠⁠⁠⁠⁠⁠Get a FREE audiobook AND support the show⁠⁠⁠⁠⁠⁠⁠⁠

• ⁠⁠⁠⁠⁠⁠⁠⁠Support the show on Patreon⁠⁠⁠⁠⁠⁠⁠⁠

• ⁠⁠⁠⁠⁠⁠⁠⁠Check out show transcripts⁠⁠⁠⁠⁠⁠⁠⁠

• ⁠⁠⁠⁠⁠⁠⁠⁠Check out our website⁠⁠⁠⁠⁠⁠⁠⁠

• ⁠⁠⁠⁠⁠⁠⁠⁠Subscribe on Apple Podcasts⁠⁠⁠⁠⁠⁠⁠⁠

• ⁠⁠⁠⁠⁠⁠⁠⁠Subscribe on Spotify⁠⁠⁠⁠⁠⁠⁠⁠

• ⁠⁠⁠⁠⁠⁠⁠⁠Subscribe on YouTube⁠⁠⁠⁠⁠⁠⁠⁠

• ⁠⁠⁠⁠⁠⁠⁠⁠Subscribe on Stitcher

Have you ever walked out of a usability session completely confident in your findings, only to ship something that quietly missed the mark?

What if the signal was there the whole time, and your brain just decided it wasn't worth logging?

This week on the Cognition Catalog, we tackle The Expectation Bias. This bias shapes what you notice before you've even decided what to think about it. Your brain has already generated a prediction before the first participant clicks a button or a teammate presents their work, and that prediction quietly shapes what registers as a signal and what gets explained away before you've made a single conscious decision about what any of it actually means.

We get into the science behind why this happens, and trace the research back to psychologist Robert Rosenthal's work in the early 1960s. His experiments, including the landmark Pygmalion in the Classroom study with Lenore Jacobson, showed that expectations don't just color our perceptions; they can actually change outcomes. That's a sobering thought when you consider how many design decisions are built on research we assumed was neutral.

We also dig into where this plays out on real teams: in usability sessions where hesitations get logged as "minor," in design reviews where leadership-championed features get a generous read while quietly doubted projects get interrogated at every turn, and in how we evaluate colleagues whose reputations have already done the evaluating for us. If any of that sounds familiar, this episode offers five concrete habits to help you catch the filter before it's already done its job. Give it a listen.

Topics:

• 00:00 - Perception is prediction

• 02:04 - A UX research cautionary tale

• 03:23 - Defining expectation bias

• 03:42 - Prediction errors explained

• 04:31 - Pygmalion effect origins

• 06:03 - Expectation vs confirmation

• 06:30 - How it warps team decisions

• 08:31 - Habits to reduce bias

• 10:47 - Wrap up and next steps

—
Thanks for listening! We hope you dug today’s episode. If you liked what you heard, be sure to like and subscribe wherever you listen to podcasts! And if you really enjoyed today’s episode, why don’t you leave a five-star review? Or tell some friends! It will help us out a ton.
If you haven’t already, sign up for our email list. We won’t spam you. Pinky swear.
• Get a FREE audiobook AND support the show
• Support the show on Patreon
• Check out show transcripts
• Check out our website
• Subscribe on Apple Podcasts
• Subscribe on Spotify
• Subscribe on YouTube
• Subscribe on Stitcher