What this programme is about
A PhD Artificial Intelligence is built around a question that does not yet have a satisfactory answer, not around completing a fixed list of AI modules.
Research might focus on machine learning, language, computer vision, reasoning, robotics, safety or another specialised area.
Most of your progress comes from reading current research, designing experiments and learning from results that often fail to support the first idea.
Strong mathematical and programming foundations are essential, but novelty is the real challenge.
Responsible research also matters. Performance gains have to be considered alongside robustness, bias, privacy and the consequences of deployment.
Choose potential supervisors and research groups carefully. Their expertise, compute resources, data access and research culture can shape the entire doctorate.
Inside the curriculum
Skills you'll build
Where this can take you
What studying this programme is like
Early doctoral work usually involves a deep literature review that identifies a narrow gap worth investigating.
You may build new models, learning methods, evaluation frameworks or theoretical results.
Experiments require careful baselines, reproducibility and enough evidence to show that an apparent improvement is meaningful.
Depending on the topic, your work may involve large-scale computing, specialised datasets or collaboration with another discipline.
Publishing and presenting research become part of the training as your findings mature.
The dissertation must ultimately show an original contribution that extends what the field already knows.
Is this likely to suit you?
Good fit if you...
- You already have advanced AI or machine-learning foundations.
- You enjoy open-ended research.
- You are comfortable with mathematics and programming.
- You want to create new knowledge rather than only apply existing tools.
- You can persist through uncertain research results.
Think twice if you...
- You mainly want practical AI skills for an industry role.
- You dislike long-term independent research.
- You are not prepared for substantial mathematics and coding.
- You expect a PhD to follow a predictable taught-course structure.