About the Project
Supervisor: Dr Han Wu
Large Language Models memorise their training data and can reproduce personal information, copyrighted content and proprietary intellectual property. This PhD project develops privacy technologies to detect, control and remove such memorisation, from individual models to multi-agent AI systems, helping to build AI that respects both privacy and creators' rights.
Large Language Models (LLMs) and the AI agents built on them are transforming how we search, write and code. Yet their power comes with a risk: they memorise their training data and can reproduce personal data, copyrighted content or proprietary intellectual property (IP). These issues are central to data protection law and ongoing copyright disputes over AI.
This project takes a privacy-centred view of memorisation, using privacy technologies to detect, control and remove unauthorised information in LLMs.
Possible directions include:
- auditing memorisation: privacy attacks that reveal whether a model has memorised, or was trained on, specific personal or copyrighted data
- mitigating memorisation: machine unlearning and privacy-preserving training that remove or limit unwanted data influence with verifiable guarantees
- multi-agent systems: tracing how data and IP flow across the AI supply chain of models, datasets, tools and agents, drawing on ideas such as AI bills of materials to make these systems transparent and accountable.
The intended outcome is practical methods for auditing and safeguarding LLMs, supporting trustworthy AI that complies with data protection and copyright law.
You will join the cyberPUNK Lab led by Dr Han Wu in the Cyber Security group, working with an interdisciplinary team spanning law, finance, human-computer interaction and industry experts in AI governance. The exact focus can be shaped around your interests.
Entry requirements
You must have a UK 2:1 honours degree or its international equivalent, in one of the following:
- computer science
- Artificial Intelligence
- or related disciplines
Strong experience in machine learning and Python programming is desirable.
Fees and funding
We offer a range of funding opportunities for both UK and international students. Horizon Europe fee waivers automatically cover the difference between overseas and UK fees for qualifying students.
Competition-based Presidential Bursaries from the University cover the difference between overseas and UK fees for top-ranked applicants.
Competition-based studentships offered by our schools typically cover UK-level tuition fees and a stipend for living costs for top-ranked applicants.
Funding will be awarded on a rolling basis, so apply early for the best opportunity to be considered.
For more information, please visit our postgraduate research funding pages.
How to apply
You need to:
- choose programme type (Research), 2027/28, Faculty of Engineering and Physical Sciences
- select Full time or Part time
- choose the relevant PhD Electronic & Electrical Engineering (7092)
- add name of the supervisor in section 2
Applications should include:
- a personal statement
- your CV (resumé)
- 2 academic references
- degree transcripts to date
Contact us
Faculty of Engineering and Physical Sciences
If you have a general question, doctoralcollege-admissions@soton.ac.uk.
Project leader
For an initial conversation, h.wu@soton.ac.uk.
The School of Electronics & Computer Science is committed to promoting equality, diversity inclusivity as demonstrated by our Athena SWAN award. We welcome all applicants regardless of their gender, ethnicity, disability, sexual orientation or age, and will give full consideration to applicants seeking flexible working patterns and those who have taken a career break. The University has a generous maternity policy, onsite childcare facilities, and offers a range of benefits to help ensure employees' well-being and work-life balance. The University of Southampton is committed to sustainability and has been awarded the Platinum EcoAward.