LLMs learn from data, but not always the right data. Because their training lacks cultural nuance and diversity, the outputs are often inequitable. Our team asked:
What if AI could learn directly from real people sharing real cultural experiences?
We wanted to build a product that:
Centers genuine human voices.
Encourages cultural exchange.
Provides a consent-driven, ethical pipeline for training more inclusive AI.
Focus
As the UX Designer, I led the charge in shaping the interaction flow to support real human connection. Emphasizing mutual understanding, not just data collection. While some team members initially leaned toward building an AI-driven language tutor, I advocated for aligning the product with the original design prompt: staying human.
I focused on:
Shaping a product vision rooted in empathy and shared learning.
Designing interaction flows that foster meaningful conversation.
Ensuring the product's value is based on shining light on under-represented groups and cultural exchange.
1. Discovery Stage
Early in the hackathon, I helped steer a pivotal reframing of the team’s idea. Initial concepts leaned heavily into automation and AI dominance, but I pushed us to revisit the design prompt. We also interviewed our software engineer friends and colleagues that specializes in AI and ML.
Our user research highlighted a critical pain point:
The lack of cross-cultural input in AI training perpetuates biased systems.
We pivoted toward a human-first platform where people from diverse backgrounds could share perspectives and stories. While optionally consenting to have that input inform better, more ethical AI.
2. Define and Design
I designed experience flows with two key goals:
Make cross-cultural exchange feel authentic and inviting.
Clearly communicate why the user’s voice matters, and how their contribution could reshape future AI systems.
My interaction design centered around a conversational UX pattern, where participants could explore different cultures, contribute their own insights, and have control over how their data was used.
3. Team Collaboration & Advocacy
Within a tight 48-hour sprint, I collaborated with UX/UI teammates across time zones. I also navigated internal disagreement about product direction. Using research findings and the core design prompt to advocate for a more values-aligned, human-centered solution.
Outcome
By the end of the hackathon, we had:
A validated product concept with clear UX value.
Interactive wireframes showcasing how users engage, share, and give informed consent.
A re-centered product vision that better aligned with the original challenge.
An ethical approach and cross-cultural focus reinforcing that empathy is a design skill.
🌱 what i learned 🌱
Design advocacy matters. I learned how to diplomatically challenge ideas that don't align with user or project goals, and how to back that up with research and clarity.
Great UX is about clarity + consent. Designing for informed user participation made me think more deeply about data ethics and transparency.
Constraints fuel creativity. This experience showed me how much thoughtful design can be done in a short timeframe when there’s alignment and intention.
