What TikTok shows, and what people watch
Donated TikTok histories linked to surveys: health content and misinformation, served and watched.
Research and projects
I study what social media feeds show people about health and the brain, and what people actually watch. My work falls into four parts.
“The brain” in public
How claims about “the brain”, ADHD and neurotechnology are framed online, and what neuroethics can learn from it.
Radicalization and misinformation
How movements build their narratives and how language changes as people move toward extreme positions.
Language models and how people get information
How large language models change clinical communication and the way people look up information.
Projects
HARMONY: Health-Related (Mis-)Information on Social Media and Its Impact on Young Adults
TikTok uses highly individualised content recommendation systems powered by artificial intelligence. While these systems boost engagement, concerns arise about their impact on young adults, particularly around mental health and exposure to health (mis-)information. HARMONY investigates how, when, and why young adults encounter health-related (mis-)information on TikTok: it collects data on users' interactions with and perceptions of TikTok's recommendation systems, analyses health-related videos, runs experiments with digital twins, and uses focus groups to learn how users judge the reliability of health content. The findings are meant to inform regulatory discussions about vulnerable groups such as young adults. An IGSSE International Project Team of TU Munich and Eindhoven University of Technology (principal investigators Marcello Ienca and Lily Frank), with the Stanford Social Media Lab and UNICEF.
TikTalks I
July 2024 – February 2026
How TikTok's recommendations build personal feeds for young adults (18–24) in Bavaria, with a focus on polarizing, radical and extremist content. Participants answered a survey, used a newly created TikTok study account for seven days, and could donate their own TikTok data; focus groups and workshops added their own accounts of what they saw. Responsible Technology Hub and TUM, funded by the Bavarian State Ministry for Family, Labour and Social Affairs under its radicalization-prevention programme. Role: project manager.
TikTalks II
2026–2027
Takes the data and results of TikTalks I to researchers, practitioners and the public: an expert workshop (held June 2026, jointly with HARMONY) and an international conference in 2027. Same partners and funder. Role: project manager.
Global Comparative Study on Censorship in LLM-Based Chatbots
September 2025 – September 2026
As large language models increasingly shape access to information, their answers to politically, ethically and socially sensitive questions raise concerns about freedom of expression, digital rights and epistemic justice. The project compares how chatbots developed under different political and regulatory conditions handle such questions, looking both for outright refusals and for subtler strategies such as evasion, distortion or omission. The current study focuses on contested historical events. TUM and Imperial College London (principal investigator Marcello Ienca, partner Giorgio Gilestro), funded by the TUM Global Incentive Fund. Role: doctoral researcher and project coordinator.