PhD Studentship: Longitudinal, Multilingual, and Multi-format Investigation and Detection of LLM-Generated Disinformation

University of Sheffield - School of Computer Science

Disinformation is now a universal problem across nations and languages, with the most recent Global Risks Report from the World Economic Forum identifying mis/disinformation as the top short/mid-term risk to society (WEF, 2025). The increasing advance of large language models (LLMs) has led to a rapid rise in LLM misuse by malicious actors, for the purposes of low-cost generation of fake news and propaganda content, i.e. the spread of AI-generated disinformation at scale (Xu et al., 2023). An extensive assessment of LLMs’ capabilities of generating high-quality disinformation is needed to understand its potential harms and propose actions to counter it.

This PhD project aims to answer the following research questions:

• Which are the best red-teaming techniques to fool LLMs into generating high-quality disinformation? Which are the linguistic and temporal impacts in the quality of LLM-generated disinformation?
• How can the quality of LLM-generated disinformation be assessed? How can hallucinations be differentiated from misleading content?
• How can LLMs be leveraged to effectively detect LLM-generated disinformation also providing explanations?

Supervisors:

Professor Kalina Bontcheva is Professor of Text Analytics and head of the GATE team part of the NLP group at The University of Sheffield. She is a world-leading researcher in Natural Language Processing. She has been a leading developer of the world-recognised NLP analytics infrastructure (GATE) since 1999. Her main interests are NLP methods for social media analysis (including online abuse and disinformation analysis), software infrastructures for NLP, information extraction, natural language generation, and text summarisation. Professor Bontcheva is the coordinator of FP7 PHEME project, scientific coordinator for H2020 WeVerify project, USFD PI for the EDMO Ireland. Currently, scientific coordinator and Sheffield PI for the HE vera.ai project and HE SoBigData++. https://sheffield.ac.uk/cs/people/academic/kalina-bontcheva

Dr Carolina Scarton is a Senior Lecturer in Natural Language Processing part of both the GATE team and the NLP group at the University of Sheffield. She is an established researcher in NLP, with over 15 years of experience in the field. Dr Scarton research interests are on social media analysis, multilinguality and multimodality, machine translation, and AI for NLP. Dr Scarton was the coordinator for the EMIF ExU project and the scientific coordinator and Sheffield PI for the HE VIGLANT project. https://sheffield.ac.uk/cs/people/academic/carolina-scarton

The School of Computer Science at the University of Sheffield is recognised for research excellence, with 99% of its research rated world-leading or internationally excellent in the 2021 Research Excellence Framework. We are rated as 8th nationally for the quality of our research environment, showing that the School of Computer Science is a vibrant and progressive place to undertake research.

How to apply:

Applications must be made directly to the University of Sheffield using the Postgraduate Online Application Form. Please name your proposed supervisor and include a CV and a cover letter in your application (2 pages max). This cover letter should describe your background (including the relevant modules taken), research interests and an overview of what you would like to focus on during the beginning of your PhD.

Funding Notes

The PhD studentship will cover standard UK home tuition fees and provide a tax-free stipend of £20,780 per annum for 3.5 years (starting in 2025/26).

Overseas students are eligible to apply but must be able to pay the difference between the UK and overseas tuition fees via additional funding or self-funding.