Student Assistant (m/f/d) – Self-Censorship in LLM-Based Chatbots (TUM-GIF)

02.12.2025, Studentische Hilfskräfte, Praktikantenstellen, Studienarbeiten

Student Assistant (m/f/d) – Self-Censorship in LLM-Based Chatbots (TUM-GIF)

Start: as soon as possible
Duration: initially 6 months (extension possible)
Hours: 5–8 hours per week
Location: Professorship of Ethics of AI and Neuroscience, Institute for History and Ethics of Medicine, TUM (hybrid/remote possible)

About the project

The project Self-Censorship in LLM-Based Chatbots (TUM × Imperial College London; GIF) investigates how large language models (LLMs) handle politically, ethically, and culturally sensitive topics across different regions of the world. Using a standardized prompt framework and NLP techniques, we evaluate the degree, form, and bypass ability of self-censorship in major LLM models.

Your role

• Assisting with data collection via OpenRouter and comparable APIs
• Running and documenting chatbot response tests
• Supporting analysis using Python or R (SBERT, MoralBERT, NER pipelines)
• Organizing and cleaning datasets for further research use
• Preparing well-structured documentation and reproducible scripts on GitHub
• Collaborating with international researchers on data organization and interpretation

What you bring

• Enrolled in a Master’s (or advanced Bachelor’s) program in Computational Linguistics, Computer Science, Data Science, or related field
• Experience with LLMs and NLP frameworks (e.g., Hugging Face)
• Familiarity with OpenRouter API or comparable LLM interfaces
• Confident working with VS Code, GitHub, and modern collaborative workflows
• Solid programming skills in Python
• Clear, well-structured scientific communication in English (spoken & written)
• Interest in interdisciplinary research at the intersection of AI ethics, language models, and digital society
• Bonus: previous experience with collaborative research projects, preprints, or agentic AI tools

What we offer

• Active participation in an international research collaboration (TUM × Imperial College London)
• Mentoring and co-authorship opportunities in scientific publications
• The option to write your Master’s thesis within the project (subject to supervision capacity)
• OpenRouter API access and technical infrastructure for experimentation
• Flexible working hours and a hybrid work setup
• Compensation according to the TUM student assistant salary table

How to apply

Please send your application as a single PDF (CV, short motivation letter [max. 1 page], current enrolment certificate, and a written sample that includes at least one self-generated graph from your coursework or a scientific project) to:
alexander.sobieska@tum.de

Subject: Student Assistant – Self-Censorship LLMs
Application deadline: 31 January 2026 (applications will be reviewed on a rolling basis; the position may be filled earlier)

For questions, please contact:

• Alexander Sobieska, MSc (alexander.sobieska@tum.de)
• Office Team: office.ethics@mh.tum.de | +49 89 4140-4041

Institutional Details

Chair of Ethics of AI and Neuroscience
Institute for History and Ethics of Medicine
Technical University of Munich
Ismaninger Str. 22, 81675 Munich
www.get.med.tum.de / www.tum.de

Data Protection Notice

As part of your application for a position at the Technical University of Munich (TUM), you provide personal data. Please note our data protection information pursuant to Art. 13 General Data Protection Regulation (GDPR) on the collection and processing of personal data in connection with your application: https://portal.mytum.de/kompass/datenschutz/Bewerbung/. By submitting your application, you confirm that you have taken note of TUM’s data protection information.

Applicants with disabilities will be given preference if equally qualified.

Kontakt: alexander.sobieska@tum.de