AI Trust & Usage Research
A qualitative research study on how young adults trust AI for financial decisions.
I led a mixed-methods study — 15 interviews with 18–29 year-olds — to understand trust barriers to AI-assisted financial decisions, then synthesized findings into a phased adoption roadmap for our financial partner.
Context
As part of Duke's DesignTK 520/521 studio, we partnered with the Fidelity Center for Applied Technology to explore AI trust and usage. AI adoption is rapidly evolving, yet trust levels vary significantly. We identified a specific gap: less than 50% of low-income, young (18–29), or Hispanic adults are currently investing in stocks. We hypothesized that the primary barriers to budgeting and investing are knowledge and confidence, rather than a lack of funds.
Initially, we aimed to understand how AI could bridge literacy gaps. We reframed the goal to explore the broader emotional landscape — how young adults feel about relying on AI for high-stakes decisions across school, work, and finances.
Research
Screening
- Age 18–29
- U.S. based
- Familiarity with AI tools
Participants
- 15 total
- Interest in finances / fintech
Method
- Tetra/Respondent online
- In-person interviews
We conducted a mix of 15 in-person and online interviews via Tetra/Respondent with U.S.-based adults aged 18–29 with existing familiarity with AI tools, then used Miro to cluster findings into actionable insights for our financial partner.
AI can advise, not decide
- “Make the investment decision yourself, but use everything that ChatGPT can offer you to learn and inform that decision” — Tetra Interviewee #2
- “I can always ask [AI for advice], but I just would not rely on it completely” — Interviewee “T”
Strong desire for accuracy
- “I generally trust [AI], but if it's a higher stakes thing then I definitely want to check the sources” — Tetra Interviewee #3
- “AI hasn't always been 100% accurate...I trust it but it's not going to be the only thing I look at” — Tetra Interviewee #5
Insight
Midway through, the project pivoted. The original framing — general trust in AI for financial decisions — gave way to a sharper, more specific insight once we dug into who was actually struggling most.
Fall '25 focus
Young adults lack trust and confidence in using AI for high-stakes financial decisions.
Spring '26 focus
Understand how young adults with ADHD interact with finance and financial platforms to decrease stress around wealth management.
Meet Sam, 22
Sam, 22
Sales Representative · Chicago, IL · Recently graduated
“I just want to stop feeling behind with my money — I know I'm spending too much, I just can't seem to fix it.”
Background
Sam just finished college and landed his first sales job in Chicago — fully financially independent for the first time, and finding it harder than expected.
Pain points
- Time blindness — notices bills, forgets to act until it's too late
- Executive dysfunction — reviewing finances feels like too much to start
- Impulse spending — small purchases add up without him noticing
Goals
- Build better habits without feeling overwhelmed
- Make intentional spending decisions
- Feel in control of his finances
How might we help young adults with ADHD manage their finances in low-effort moments that reduce executive overload and time-blind avoidance?
Design
Notification dread cycle
Users with ADHD often avoid financial apps due to overwhelming or poorly-timed notifications, creating a cycle of avoidance.
Gamification for dopamine
Reward-driven mechanics align with how ADHD brains seek motivation, making habit formation more accessible.
Adaptive timing
Flexible, context-aware nudges beat fixed, high-frequency reminders for time-blind users.
Context-dependent push notifications
Signals like biometrics, location, and calendar completion surface a nudge exactly when the user is ready — not at a fixed time.
Visual aesthetic
A plant-based metaphor makes financial health tangible — savings and bill coverage become visible, emotional progress worth tending to.
Gamified habit building
Reward-driven mechanics create immediate feedback that aligns with how ADHD brains seek motivation.
Streamlined onboarding
Home dashboard & plant view
Plant growth journey
Context-aware notifications
Outcome
We delivered a three-phase roadmap to our financial partner, then carried the research forward into a working prototype the following semester.
Phase 1
Human-first brandingPosition AI as a support system for human advisors, emphasizing empathy and reliability.
Phase 2
Backend deploymentUse AI as a productivity booster for internal teams rather than a direct-to-customer interface.
Phase 3
Clarify & Learn featureAn AI feature that explains financial terms with full transparency, citations, and verified-by-expert badges.
Reflection
What we learned
- Subtle, timing-based prompts help users start tasks without adding cognitive pressure
- The plant analogy makes progress feel visible and emotional
Next steps
- Move toward personalized, predictive support using biometric or behavioral data
- Conduct more user testing on the developed prototype
Next project
Shipt Search Filters →