Interview and survey data, thematic analysis, AI-integrated data analysis.
TEA APP DAta Ethics and recommendations
ROLE: Researcher / Data Analyst / Policy Designer
TIMELINE: 2025 (Aug - Dec)
SKILLS: Speculative Design, Critical UX, Interaction Design, Qualitative Research, Visual Analysis
IN COLLABORATION WITH: Elizabeth Travia and Ely Cariveau, Information Ethics and Policy with Prof. Chris Carruth
Navigating the modern dating scene increasingly means navigating digital risk, where apps designed to support communication and connection can also introduce fears about privacy breaches, misinformation, and unwanted exposure. The Tea App is a clear example of this complicated landscape because it encourages women to share experiences about men they are dating or have dated in order to promote community safety, warn others, or highlight positive interactions. What began as a space for empowerment and mutual protection has quickly become the subject of ethical, social, and political debate, largely because of concerns about doxxing, unverified claims, misuse of personal data, and the potential for harm. How can we conduct user research to best understand how people feel about the Tea App and their individual data privacy, to speculate on a new competing platform?
Research
Research Question: how can the Tea App violate or protect user privacy while also attempting to protect users’ physical safety? What is considered public safety/intentional doxxing?
Stakeholders: Application users/people mentioned, application developers, privacy regulators.
Relevance: Highlights imperfections in online privacy rules and laws. Recognizes negative impact of social media on others.
Target Audience: Tea App users and developers who need a clear definition on the ‘line’ between intentional, necessary public safety (this person is a sex offender) versus intentional doxxing.
METHODOLOGIES
Selection: Gathered comments from online forums, filtering by recent, old, and most popular
Manual Coding: Generated 3 codes per comment, ranking each by relevance for a primary, secondary, and tertiary code
AI Assisted Thematic Analysis: Original data + codes were inputted into ChatGPT with the prompt:
‘use this data from reddit and similar online forums and create a set of themes using traditional thematic analysis.’
Limitations and Bias: The scope of our manual comment collection is limited, and may introduce algorithmic bias based on what comments were visible to us.
AI integrated Thematic Analysis
Final Policy Recommendation
Policy Recommendation
The Tea Application aims to create a safe environment for women to share about their experiences with men in their lives safely without fear. This is not a perfect system.
Current Failures with Tea:
Doxxing/Malintent: The idea of ‘tea’ promotes gossip and opens the floodgates for lying about individuals.
Data privacy: Tea experienced a massive data leak which lead to the addresses and drivers’ licenses of many users to be leaked online.
Can contribute to violent reactions from men (who manage to get on the app)
Policy for a new women’s safety app:
Uses PUBLIC DATA - not user added info.
Sex offender / criminal regestries.
Allows users to privately DM each other about specific people they’re concerned about. Removes the doxxing capability and fear of retaliation.
Provides resources for reporting to the police and additional women’s safety tools.
