Research: People and AI

About This Research

AI tools are becoming part of how people work, think, and relate to each other. But most of what we hear about AI adoption is either hype or fear. There's remarkably little grounded research into the actual lived experience of people who are integrating these tools into their daily routines.

This project looks beyond the hot takes. Through in-depth interviews with people across roles, industries, and levels of AI experience, I'm building a rich, evidence-based picture of how AI adoption actually works: what drives it, what it changes, what hidden costs it imposes, and what it enables.

Methods

The study uses two complementary data-gathering techniques. First, a short pre-interview survey collects organizational and demographic context, along with initial descriptions of how participants and their organizations are using AI. Then, a 30-minute semi-structured interview digs deeper into the stories behind those survey responses.

The interviews follow behavioral interviewing techniques, asking for specific stories and concrete examples rather than general opinions. Topics cover the full arc of AI adoption: from first use through current patterns, including successes and failures, changes to work and personal activities, trust calibration, social dynamics around disclosure, and cognitive and emotional effects. All sessions are recorded, transcribed, and de-identified for analysis.

Goals

The goal is to gather, synthesize, and share publicly the results of this research. Outputs will include a synthesis of key findings and themes, a framework for understanding AI adoption patterns, a typology of how work practices are changing, and a map of the social norms emerging around AI use and disclosure. All findings will be published here as the research progresses, making this a living, evolving body of work rather than a single report.

To that end, everything on this site is published under a Creative Commons BY-NC-SA 4.0 license. The study materials, analytical methods, codebook, and findings are all freely available to use, adapt, and build on with attribution. Public research about how people actually use AI should itself be public, and the CC license makes that concrete: other researchers, designers, and teams working on these same questions can take what's here and run with it.

Emerging Themes

As sessions accumulate, patterns are starting to emerge. So far, 34 themes have been identified across 5 categories, covering everything from adoption patterns and trust calibration to concerns about skill erosion and job displacement.

This project is open source under a Creative Commons BY-NC-SA 4.0 license. In that spirit, I built a lightweight theme explorer so you can dig into the codebook, browse themes by category, and follow them back to the sessions where they appeared. Try it out.

Become a Participant

I'd love to talk with you for 30-45 minutes about your experiences with AI. Whether you're a power user, a cautious experimenter, or somewhere in between, I want to hear from you.

This is a personal research project. There are no sponsors and no corporate agenda, just genuine curiosity from someone who has spent more than 20 years studying how people adapt to new technology.

What's in it for you: a transcript and analysis of your session, plus access to upcoming research readouts and highlights. All results are de-identified. Your personal information won't be shared with anyone.

Session Writeups

Interim Findings

(April 23, 2026) I've analyzed 12 participants' session data and written up an in-flight description of the findings so far. This isn't a full report. Findings and conclusions will most definitely change as I collect more data. In the spirit of openness, I'm posting the interim findings here.

(AI disclosure: I wrote the description of each finding. Then I ran my prose through Claude to correct spelling and grammar errors.)

Interim Findings (v1, April 23, 2026)

Under the Hood

Glad you scrolled this far. Here's the part I'm genuinely excited about: I designed the entire analysis pipeline for this project around AI-assisted workflows, and the speed is sort of absurd. A raw interview recording becomes a de-identified, coded, and published session writeup with pull quotes, theme annotations, and a full analytical narrative in under an hour. That turnaround would have been unthinkable two years ago.

The study design, the analysis workflow, and the tooling are all still evolving as I run sessions and learn what works. This is a living project, not a frozen archive. I'm iterating on the process in the open because I think the methods are as interesting as the findings, and because other researchers might find them useful.

Everything here is published under a Creative Commons BY-NC-SA 4.0 license. Creative Commons is a nonprofit that provides free, standardized licenses for sharing creative and academic work. BY-NC-SA means you can use, adapt, and build on anything here as long as you give attribution, keep it non-commercial, and share your work under the same terms. The study materials, the analytical process, the findings: all of it is yours to learn from and build on.

AI Use Disclosure

I used AI to analyze the data collected via interviews and surveys. How?

• I took notes after each session.
• I fed those notes to several AIs, along with the moderator guide, project proposal, session transcript, the participant's survey responses, and a codebook of tags and themes I've been iterating as I collect data.
• I prompted each to write a background, findings, and emerging themes section.
• Then I iterated on each AI's draft, challenging the AI where appropriate and removing what I'm euphemistically calling "hallucinatory content" :-).
• I collected each AI's drafts, added them to the project I've set up in Claude Cowork, and prompted it to draft the background, findings, and emerging themes section, pushing back as appropriate.
• Then I edited the content, because "human in the loop" means "I have final edit." At least to me it does.
• I then published each session writeup.

There's a bit more to it, but I'm trying to keep this short. Reach out if you want to talk about my AI-assisted workflow, which I'm still evolving as I go.