Students from our Law and Technology LLM and postgraduate courses in Computer Science and Engineering collaborated in a series of Multidisciplinary Research Cafes, where they debated technical and legal solutions to real algorithmic transparency, bias, and data scraping.
This project, made possible by support from the Quintin Hogg Trust, aimed at enhancing multidisciplinary communication and skills.
Hear from Data Science and Analytics MSc student Mahnoor Bilal about what made this project so rewarding.
Can an algorithm be sued for a misdiagnosis?
This was the question at the heart of the first Multidisciplinary Research Café, an initiative bringing together students from the Law and Technology LLM alongside students from Computer Science and Engineering.
The project brings us together to collaboratively design AI systems that are not only technically robust, but also explainable, inclusive, and compliant with legal frameworks such as the GDPR and HIPAA. The first challenge: reimagine AI in healthcare — a field where algorithmic decisions carry life-altering consequences.
As to whether an algorithm could be sued? Not yet. But in an era where AI systems influence who receives a loan, a transplant, or an early cancer screening, the question is no longer theoretical — it’s legal, ethical, and deeply urgent. As systems grow more complex, so do the harms they can quietly encode.
Working together
We were divided into mixed-discipline teams of students from Law and Data Science to explore how AI systems could be responsibly deployed in healthcare. My group focused on algorithmic fairness — particularly how the underrepresentation of ethnic groups in medical datasets can lead to skewed outcomes in diagnosis and treatment. We examined tools like SHAP (SHapley Additive Explanations) to interpret model predictions and discussed how to embed principles like data protection, transparency, and consent into the design process itself.
The law students brought legal precedents, regulatory frameworks, and ethical red flags to the table. Those of us from data science worked through the mechanics: building interpretability into models, identifying features that may act as proxies for race or class, and questioning what “accuracy” means when it risks reproducing structural inequality.
A moment of ethical friction
One moment that stayed with me came during a debate over postcode data. In our dataset, postcode was a strong predictor for certain health outcomes. Technically, including it boosted the model’s accuracy. But we knew that postcode often correlates with ethnicity and socioeconomic status. Should we use it, knowing it sharpens prediction? Or exclude it, to avoid reinforcing historical disparities? From a data science view, it was a rational feature. From a legal perspective, it bordered on indirect discrimination. That tension — between performance and principle — wasn’t just academic. It was alive.
What I learned
What surprised me most was the conversation. Explaining machine learning techniques to law students, and seeing them reframe those techniques through the lens of legal responsibility, made one thing clear: we weren’t just collaborating or simplifying for each other. We were finding shared language for shared responsibility.
As a data science student, I’ve studied fairness and bias in modules and textbooks. But this experience showed me what it means to design fairness into a real system. That shift in thinking has already shaped how I approach my coursework, particularly in modules related to AI governance, machine learning ethics, and data systems.
What's next: from ideas to implementation
In the next sessions, we’ll move from discussion to design — translating our frameworks into visual architectures and working guides. Our goal is to co-create a resource that unites technical strategies with legal guardrails — something that could one day inform how practitioners in law and healthcare approach algorithmic fairness.
But we’re also building something harder to define: a way of thinking that moves beyond compliance, toward care. A way of building systems that don’t just work — but work justly.
This initiative was made possible thanks to the vision and leadership of Dr. Diana Sancho and Dr. Anastassia Angelopoulou, whose work continues to bridge the gap between law and technology.
A transformative experience
For any student working at the intersection of technology and society, experiences like this are transformative. It challenged me to think beyond metrics and models — to consider who our systems serve, and who they might leave behind. Because in the end, algorithms don’t live in code. They live in people’s lives.
And if we want to shape the future of AI, we have to shape it with them in mind.
About Westminster
As one of the most diverse universities in the UK, we are a global university with London energy, with more than 19,000 students from 169 countries. To find out more about our Law and Technology LLM and Computer Science and Engineering courses, visit our website.
- Can an algorithm be sued? Lessons from the Law & Technology frontier - July 9, 2025
- A week in the life of a Psychology student: Saumyaand Anandi - March 11, 2025
- Why do we need new and diverse approaches to police education? - August 9, 2024