Dr Paul Röttger
About
Dr Paul Röttger is a Departmental Lecturer at the Oxford Internet Institute (OII) and an Associate Member of Nuffield College at the University of Oxford. His work is located at the intersection of natural language processing (NLP) and the social sciences.
In his current research, Dr Röttger focuses on evaluating and improving the safety of large language models (LLMs) as well as studying their societal impacts. For his recent work in this area, he won an Outstanding Paper Award at ACL, the top NLP conference, and a Best Paper Award at NeurIPS, the top AI conference.
Before joining the OII as Lecturer, Dr Röttger was a Postdoctoral Researcher at Bocconi University, working with Professor Dirk Hovy and the MilaNLP lab. Dr Röttger also holds a DPhil from the OII, where he was supervised by Professor Helen Margetts and Professor Janet Pierrehumbert. During his DPhil, Dr Röttger co-founded a start-up building AI tools for content moderation, which was acquired by another large online safety company in 2023.
Expertise
- AI safety
- AI risks
- AI evaluation
- Large language model (LLM) bias
- EU AI regulation
Selected publications
Recent media work
- When the AI says: "It could be a new beginning to have your husband killed" (Frankfurter Allgemeine, 2026)
- When AI chatbots become malicious (Tagesschau, 2026)
- What comes after Deep Seek? (Süddeutsche Zeitung, 2025)
- Why Trump's order targeting 'woke' AI may be impossible to follow (New Scientist, 2025)
- Bang with announcement (Zeit Online, 2025)
- Why the "thinking ability" of Deepseek R1 and ChatGPT o1 is a breakthrough (Der Standard, 2025)
- US tech stocks are rising again: Is the "DeepSeek shock" not so big after all? (Tagesspiegel, 2025)
- Startup firm Patronus creates diagnostic tool to catch genAI mistakes (ComputerWorld, 2024)
- What is ChatGPT-4, the latest OpenAI bot? (The Times, 2023)
- OpenAI’s red team: the experts hired to ‘break’ ChatGPT (Financial Times, 2023)
