Rednote AI Trust

Rednote AI Trust

Rebuilding user trust in AI-generated content

Rebuilding user trust in AI-generated content

Role: Lead UX Researcher & Designer

Role: Lead UX Researcher & Designer

Timeline: 03/2026–04/2026

Timeline: 03/2026–04/2026

Tools: Figma, Prototyping, Usability Testing

Tools: Figma, Prototyping, Usability Testing

The Problem

Rednote's "Ask AI" feature improved search efficiency, but created a trust problem: users shifted from active explorers to passive recipients. Three core UX issues:

Inconsistent Explainability

No User Control

Invisible Feedback Loop

Research

Methodology: Product analysis + comparative study + decision matrix

Compared Rednote Ask AI against Amazon "Customers say" and TripAdvisor AI Summary across interpretability, user control, and traceability dimensions.

Decision matrix identified Explainability and User Control as high-priority, high-feasibility intervention points — the foundation for rebuilding trust without disrupting existing browsing habits.


① Pre-generation Control

① Pre-generation Control

② In-generation Switching

② In-generation Switching

③ Sentence-level Evidence Drawer

③ Sentence-level Evidence Drawer

④ Closed-loop Feedback

④ Closed-loop Feedback

Result

Task-based usability walkthrough with 4 participants. All completed all tasks, most within 1 minute.


Participants highlighted the "Real users" and "No promos" filters as significantly reducing cognitive load. Sentence-level evidence mapping was identified as the key trust-building element.

4

4

Participants

Participants

4.25/5

4.25/5

Efficiency

Efficiency

3.75/5

3.75/5

Trust

Trust

Contact

Wechat

Email

Whatsapp

Tel

Create a free website with Framer, the website builder loved by startups, designers and agencies.